199 64 3MB
English Pages 217 Year 2009
Command and Control: The Sociotechnical Perspective
Guy H. Walker, Neville A. Stanton, Paul M. Salmon and Daniel P. Jenkins
COMMAND AND CONTROL: THE SOcIOTEcHNIcAL PERspEcTIVE
Human Factors in Defence Series Editors: Dr Don Harris, Cranfield University, UK Professor Neville Stanton, Brunel University, UK Professor Eduardo Salas, University of Central Florida, USA Human factors is key to enabling today’s armed forces to implement their vision to ‘produce battle-winning people and equipment that are fit for the challenge of today, ready for the tasks of tomorrow and capable of building for the future’ (source: UK MoD). Modern armed forces fulfil a wider variety of roles than ever before. In addition to defending sovereign territory and prosecuting armed conflicts, military personnel are engaged in homeland defence and in undertaking peacekeeping operations and delivering humanitarian aid right across the world. This requires top class personnel, trained to the highest standards in the use of first class equipment. The military has long recognised that good human factors is essential if these aims are to be achieved. The defence sector is far and away the largest employer of human factors personnel across the globe and is the largest funder of basic and applied research. Much of this research is applicable to a wide audience, not just the military; this series aims to give readers access to some of this high quality work. Ashgate’s Human Factors in Defence series comprises of specially commissioned books from internationally recognised experts in the field. They provide in-depth, authoritative accounts of key human factors issues being addressed by the defence industry across the world.
Command and Control: The Sociotechnical Perspective
Guy H. Walker, Heriot-Watt University Neville A. Stanton, University of Southampton Paul M. Salmon, Monash University, Australia & Daniel P. Jenkins, Sociotechnic Solutions, UK
© Guy H. Walker, Neville A. Stanton, Paul M. Salmon and Daniel P. Jenkins 2009 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior permission of the publisher. Guy H. Walker, Neville A. Stanton, Paul M. Salmon and Daniel P. Jenkins have asserted their right under the Copyright, Designs and Patents Act, 1988, to be identified as the authors of this work. Published by Ashgate Publishing Limited Ashgate Publishing Company Wey Court East Suite 420 Union Road 101 Cherry Street Farnham Burlington Surrey, GU9 7PT VT 05401-4405 England USA www.ashgate.com British Library Cataloguing in Publication Data Command and control : the sociotechnical perspective. -(Human factors in defence) 1. Command and control systems. 2. Military art and science--Technological innovations. 3. Human-computer interaction. I. Series II. Walker, Guy. 355.3'3041-dc22
ISBN: 978-0-7546-7265-4 (hbk) 978-0-7546-9191-4 (ebk.IV)
Library of Congress Cataloging-in-Publication Data Command and control : the sociotechnical perspective / by Guy Walker ... [et al.]. p. cm. -- (Human factors in defence) Includes bibliographical references and index. ISBN 978-0-7546-7265-4 (hardback) -- ISBN 978-0-7546-9191-4 (ebook) 1. Command and control systems--Effect of technology on. 2. Command and control systems--Technological innovations. 3. Command and control systems--Evaluation. 4. Communications, Military. 5. Technology--Social aspects. I. Walker, Guy. UB212.C637 2009 355.3'3041--dc22 2009026212
Contents List of Figures List of Tables Acknowledgements About the Authors
ix xiii xv xvii
1
Introduction Network Enabled Capability Where Did the Humans Go? Sociotechnical Theory Parallel Universes Matching Approaches to Problems What is the Network Enabling? An Interconnected Approach to Analysis From Theory to Practice Summary
1 1 3 6 7 8 8 9 9 10
2
Reconsidering the Network in NEC Aims of the Chapter Rationality and Industrial Age Thinking The Irrationality of Rationality Network Enabled Capability Main Principles of Sociotechnical Theory Responsible Autonomy Adaptability Meaningfulness of Tasks Sociotechnical Systems Import and Export Steady States Equifinality Real-Life Sociotechnical Systems Organisational Interventions The Sociotechnical Legacy Contemporary Sociotechnical Systems Theory Summary
11 11 11 14 15 16 17 18 19 20 20 21 21 22 23 24 26 29
3
Some Effects of Certain Communications Patterns on Group Performance Aims of the Chapter
31 31
vi
Command and Control: The Sociotechnical Perspective
Introduction Method Procedure Results Summary
31 37 44 45 55
4
Complexity and Human Factors Aims of the Chapter Human Factors Complexity The ‘Attribute’ View The Complex Theoretic View The Complex Systems Research View Command and Control Summary
57 57 57 58 63 66 70 76
5
Dimensions of Live-NEC Aims of the Chapter Introduction Part 1: Developing the NATO SAS-050 Model Part 2: Exploring Live-NEC Using the Enhanced NATO SAS-050 Approach Space Modelling Results: Digitally Mediated Communications Results: Agility Summary
77 77 77 81 88 90 91 99 102
6
The Human in Complex Dynamic Systems Aims of the Chapter Introduction Emergence Sensitive Dependence on Initial Conditions From Typology to Taxonomy Summary
103 103 103 104 115 117 123
7
Beyond NEC Aims of the Chapter The Ultimate NEC Dimensions of Organisation Structure Beyond NEC NATO SAS-050 Derived Archetypes NEC versus Commercial Organisations NEC Benchmarked against Commercial Organisations NEC versus Civilian Command and Control NEC Benchmarked Against Civilian Command and Control
125 125 125 126 128 129 132 135 136 139
Contents
vii
NEC versus ‘Next Generation’ Organisations NEC Benchmarked Against Next Generation Organisations NEC versus Jointly Optimised Sociotechnical Organisations NEC Benchmarked against a Sociotechnical Organisation Summary
141 144 146 149 150
8
The Design of Everyday Networked Interoperable Things Aims of the Chapter The Proposition(s) The Information Age Design Evolution Design Co-evolution Open Systems Behaviour Summary
153 153 153 154 155 159 162 168
9
Conclusions Putting the Human Back into NEC Sociotechnical Principles for NEC System Design Visions of Success The Intervention of Last Resort
171 171 173 173 176
References and Bibliography Index
179 193
This page has been left blank intentionally
List of Figures Figure 2.1
The NATO SAS-050 approach space positions traditional hierarchical command and control in a three-dimensional space defined by unitary decision rights, hierarchical patterns of interaction and tight control of information 13 Figure 3.1 Leavitt’s network archetypes 33 Figure 3.2 Command and control microworld 39 Figure 3.3 Scatter plot showing the regression lines for both NEC and C2 conditions in relation to task time 47 Figure 3.4 Scatter plot showing the regression lines for both NEC and C2 conditions in relation to total communications 49 Figure 3.5 Scatter plot showing the regression lines for both NEC and C2 conditions in relation to attrition score 50 Figure 3.6 Scatter plot showing the regression lines for both NEC and C2 conditions in relation to team cohesion 53 Figure 4.1 Number of articles in the mainstream, peer-reviewed human factors literature that have either the word ‘complex’ or ‘complexity’ in their title 58 Figure 4.2 Alberts and Hayes (2006) problem space provides a 3D approximation of the attribute view of complexity 62 Figure 4.3 Complexity profiles for NEC and traditional hierarchical command and control 74 Figure 5.1 The NATO SAS-050 approach space is yoked to a corresponding problem space 80 Figure 5.2 The effects of scale on the position that edge organisations and classic C2 occupy in the NATO SAS-050 approach space 84 Figure 5.3 Illustration of archetypal networks 86 Figure 5.4 Leavitt’s network archetypes, along with the classic C2 and edge organisation archetypes, anchored into the enhanced NATO SAS-050 approach space 87 Figure 5.5 Organisational centre of gravity for live-NEC case study (digital comms function) 92 Figure 5.6 Results of plotting all 34 digital comms networks into the enhanced NATO SAS-050 approach space 93 Figure 5.7 Periodogram illustrating the presence of periodic changes in the distribution of information (network density) within live-NEC 94 Figure 5.8 Spectral analysis graph illustrating the presence of periodic changes in the pattern of interaction (diameter) 95
Command and Control: The Sociotechnical Perspective
Figure 5.9 Organisational centre of gravity for the live-NEC case study (voice comms function) 96 Figure 5.10 Results of plotting all 32 voice comms networks into the enhanced NATO SAS-050 approach space 97 Figure 5.11 Spectral analysis graph illustrating the presence of periodic changes in network density 98 Figure 5.12 Spectral analysis graph illustrating the presence of periodic changes in high-status nodes 99 Figure 5.13 Visual representation of the total region(s) occupied by the digital and voice functions contained within the live-NEC case study 100 Figure 5.14 A rough order of magnitude measure of total agility shows that the voice-based communications architecture in the live-NEC case study was more agile than the digital rchitecture 101 Figure 6.1 Bar chart showing the type of data comms that are being transmitted by BDE HQ 108 Figure 6.2 Bar chart showing the type of data comms that are received by BDE HQ 109 Figure 6.3 Bar chart showing the content of voice comms transmitted and received, according to Bowers et al.’s (1998) taxonomy 111 Figure 6.4 The biological conception of NEC sees a constant throughput of information, processing, output, steady states and the ability to evolve and adapt 115 Figure 6.5 The Lorenz attractor: phase space showing the complex dynamical behaviour of a physical system 119 Figure 6.6 Multidimensional space collapsed into three leading axes via PCA 121 Figure 7.1 The spectrum of organisation types generated by the Aston Studies are a 3D cube conceptually very similar to the NATO SAS-050 approach space 127 Figure 7.2 Archetypal networks representing Classic C2 and Edge Organisation 130 Figure 7.3 NATO SAS-050 approach space populated with network archetypes and a live instance of NEC 131 Figure 7.4 NATO SAS-050 approach space populated with commercial organisations from the Aston Studies, a live instance of NEC, as well as being divided up into the empirical taxonomy offered by Pugh et al. 1969 135 Figure 7.5 Social network derived from live observation of air traffic control scenario 137 Figure 7.6 Social network derived from live observation of fire service training scenario 138
List of Figures
Figure 7.7 Figure 7.8 Figure 7.9 Figure 7.10 Figure 7.11 Figure 7.12 Figure 7.13 Figure 7.14 Figure 8.1 Figure 8.2 Figure 8.3 Figure 8.4
xi
Social network derived from live observation of police service incident response scenario 139 NATO SAS-050 approach space populated with civilian command and control organisations and a live instance of NEC 140 Alternative network topologies with their corresponding histograms showing the characteristic spread of nodes having high and low levels of connectivity 141 The small-world network archetype revealing the topological features associated with high clustering and a short average path length 143 Classic C2, Edge and terrorist organisations with their corresponding histograms showing the characteristic spread of nodes having high and low levels of connectivity 145 NATO SAS-050 approach space populated with net-enabled organisations and a live instance of NEC 146 The organisation ‘after’ redesign. A simple organisation with complex tasks 147 NATO SAS-050 approach space populated with pre- and post-sociotechnical organisation 150 Bowman’s (highly simplified) evolutionary timeline 156 Hirshhorn’s Law of Stretched ‘Equipment’ 160 Interaction pull, technology push and equipment co-evolution 163 From Industrial to information-age products 167
This page has been left blank intentionally
List of Tables Table 2.1 Table 2.2 Table 2.3 Table 3.1
The use and effectiveness of common sociotechnical measures 25 Comparison of sociotechnical contexts 28 Comparison of concepts: NEC versus Sociotechnical Systems 30 Performance characteristics of Leavitt’s (1951) network archetypes. 34 Table 3.2 Experimental participants and their roles 38 Table 3.3 NATO SAS-050 Model of Command and Control was used to design NEC and C2 command organisations with the appropriate characteristics 40 Table 3.4 Question items drawn from the Platoon Cohesion Questionnaire (Siebold and Kelly, 1988) mapped onto Hackman and Oldman’s (1980) core job characteristics to create a much simplified assessment method 43 Table 3.5 Results of analysing sociometric status in relation to NEC and C2 conditions 46 Table 3.6 Standardised beta coefficients showing the relative factor loading of the individual communications categories within a regression model that assesses their contribution to total communications 54 Table 4.1 Wider trends associated with opposite corners of the human factors problem space 63 Table 4.2 Examples of complexity theory metrics explained using Hierarchical Task Analysis as the ‘computational equivalent’ for an entity or artefact 65 Table 4.3 Task analysis methods all rely on the analysis of parts in order to understand the whole 69 Table 4.4 Matrix of ‘Approach’ versus ‘Problem’ and a simple taxonomy of resultant system behaviours 75 Table 5.1 The range of values on each of the enhanced NATO SAS-050 model axes and a composite measure of agility 100 Table 6.1 Types of emergence, the information needed in order to make a diagnosis of emergent phenomena and the associated level of difficulty in doing so 105 Table 6.2 Definition of digital communication types that occur during live-NEC case study 107 Table 6.3 Free versus ‘constrained’ digital comms that are ‘transmitted’ 108 Table 6.4 Free versus ‘constrained’ data comms that are ‘received’ 109 Table 6.5 Bowers et al. (1998) communications typology 111
xiv
Command and Control: The Sociotechnical Perspective
Table 6.6
Standardised beta coefficients show the relative contribution that each voice comms subscale makes towards total communications (for both transmit and receive events) Table 6.7 Data vs. information received and transmitted by Brigade Headquarters Table 6.8 Data vs. information received and transmitted by Brigade Headquarters Table 7.1 Attributes of jointly optimized sociotechnical systems Table 8.1 Matrix of ‘Approach’ versus ‘Problem’ Table 9.1 Sociotechnical Principles of NEC System Design Table 9.2 Large-scale project failures
112 113 114 148 157 174 176
Acknowledgements The Human Factors Integration Defence Technology Centre (HFI DTC) is a consortium of industry and academia working in cooperation on a series of defence-related projects. The consortium is led by Aerosystems International and comprises the University of Southampton, the University of Birmingham, Cranfield University, Lockheed Martin, MBDA and SEA Ltd. The consortium was recently awarded The Ergonomics Society President’s Medal for work that has made a significant contribution to original research, the development of methodology and application of knowledge within the field of ergonomics. Aerosystems International
University of Southampton
Birmingham University
Cranfield University
Dr Karen Lane
Professor Neville Stanton
Prof Chris Baber
Dr Don Harris
Linda Wells
Dr Guy Walker
Professor Bob Stone
Andy Farmilo
Kevin Bessell
Dr Daniel Jenkins
Dr Huw Gibson
Geoff Hone
Nicola Gibb
Dr Paul Salmon
Dr Robert Houghton
Jacob Mulenga
Robin Morrison
Laura Rafferty
Richard McMaster
Ian Whitworth
Dr Carol Deighton
John Huddlestone Antoinette Caird-Daley
Lockheed Martin UK
MBDA Missile Systems
Systems Engineering and Assessment (SEA) Ltd
Mick Fuchs
Steve Harmer
Dr Georgina Fletcher
Lucy Mitchell
Dr Carol Mason
Dr Anne Bruseberg
Rebecca Stewart
Chris Vance
Dr Iya Solodilova-Whiteley
David Leahy
Ben Dawson
We are grateful to DSTL who have managed the work of the consortium, in particular to Geoff Barrett, Bruce Callander, Jen Clemitson, Colin Corbridge, Roland Edwards, Alan Ellis, Jim Squire, Alison Rogers and Debbie Webb. We are also grateful to Dr John Ardis for review and comment on the work that comprises Chapters 4 and 6, and to Kevin Bessell for assistance in Chapter 5. This work from the Human Factors Integration Defence Technology Centre was part-funded by the Human Sciences Domain of the UK Ministry of Defence Scientific Research Programme. Further information on the work and people that comprise the HFI DTC can be found on www.hfidtc.com.
This page has been left blank intentionally
About the Authors Dr Guy H. Walker School of the Built Environment, Heriot-Watt University, Edinburgh, [UK] EH14 4AS [email protected] Guy Walker has a BSc Honours degree in Psychology from the University of Southampton and a PhD in Human Factors from Brunel University. His research interests are wide ranging, spanning driver behaviour and the role of feedback in vehicles, railway safety and the issue of signals passed at danger, and the application of sociotechnical systems theory to the design and evaluation of military command and control systems. Guy is the author/co-author of over forty peer reviewed journal articles and several books. This volume was produced during his time as Senior Research Fellow within the HFI DTC. Along with his colleagues in the research consortium, Guy was awarded the Ergonomics Society’s President’s Medal for the practical application of Ergonomics theory. Guy currently resides in the School of the Built Environment at Heriot-Watt University in Edinburgh, Scotland, working at the cross-disciplinary interface between engineering and people. Professor Neville A. Stanton HFI DTC, School of Civil Engineering and the Environment, University of Southampton, Southampton, [UK] SO17 1BJ. [email protected] Professor Stanton holds a Chair in Human Factors and has published over 140 international peer-reviewed journal papers and 14 books on Human Factors and Ergonomics. In 1998, he was awarded the Institution of Electrical Engineers Divisional Premium Award for a co-authored paper on Engineering Psychology and System Safety. The Ergonomics Society awarded him the President’s medal in 2008 and the Otto Edholm Medal in 2001 for his contribution to basic and applied ergonomics research. The Royal Aeronautical Society awarded him the Hodgson Medal and Bronze Award with colleagues for their work on flight-deck safety. Professor Stanton is an editor of Ergonomics and on the editorial board of Theoretical Issues in Ergonomics Science and the International Journal of Human Computer Interaction. Professor Stanton is a Fellow and Chartered Occupational Psychologist registered with The British Psychological Society, and a Fellow of The Ergonomics Society. He has a BSc in Occupational Psychology from Hull University, an MPhil in Applied Psychology from Aston University, and a PhD in Human Factors, also from Aston.
xviii
Command and Control: The Sociotechnical Perspective
Paul M. Salmon Human Factors Group, Monash University Accident Research Centre, Monash University, Victoria, Australia [3800] [email protected] Paul Salmon is a senior research fellow within the Human Factors Group at the Monash University Accident Research Centre and holds a BSc in Sports Science, an MSc in Applied Ergonomics, and a PhD in Human Factors. Paul has eight years experience of applied Human Factors research in a number of domains, including the military, aviation, and rail and road transport, and has worked on a variety of research projects in these areas. This has led to him gaining expertise in a broad range of areas, including situation awareness, human error, and the application of Human Factors methods, including human error identification, situation awareness measurement, teamwork assessment, task analysis, and cognitive task analysis methods. Paul has authored and co-authored four books and numerous peerreviewed journal articles, conference articles, and book chapters, and was recently awarded the 2007 Royal Aeronautical Society Hodgson Prize for a co-authored paper in the society’s Aeronautical Journal and, along with his colleagues from the Human Factors Integration Defence Technology Centre (HFI DTC) consortium, was awarded the Ergonomics Society’s President’s Medal in 2008. Dr Daniel P. Jenkins Sociotechnic Solutions, St Albans, Herts, [UK]AL1 2LW info@sociotechnicsolutions Dan Jenkins graduated in 2004 from Brunel University with an M.Eng (Hons) in Mechanical Engineering and Design, receiving the ‘University Prize’ for the highest academic achievement in the school. As a sponsored student, Dan finished university with over two years’ experience as a design engineer in the automotive industry. Upon graduation, Dan went to work in Japan for a major car manufacturer, facilitating the necessary design changes to launch a new model in Europe. In 2005, Dan returned to Brunel University taking up the full-time role of research fellow in the Ergonomics Research Group, working primarily on the HFI DTC project. Dan studied part time on his PhD in Human Factors and interaction design – graduating in 2008 – receiving the ‘Hamilton Prize’ for the Best Viva in the School of Engineering and Design. Both academically and within industry, Dan has always had a strong focus on Human Factors, system optimization, and design for inclusion. Dan has authored and co-authored numerous journal papers, conference articles, book chapters, and books. Dan and his colleagues on the HFI DTC project were awarded the Ergonomics Society’s President’s Medal in 2008. Dan now works as a consultant and his company is called Sociotechnic Solutions.
Chapter 1
Introduction
‘The […] military could use battlefield sensors to swiftly identify targets and bomb them. Tens of thousands of warfighters would act as a single, self-aware, coordinated organism. Better communications would let troops act swiftly and with accurate intelligence, skirting creaky hierarchies. It’d be “a revolution in military affairs unlike any seen since the Napoleonic Age.” And it wouldn’t take hundreds of thousands of troops to get a job done—that kind of “massing of forces” would be replaced by information management. […] Computer networks and the efficient flow of information would turn [the] chain saw of a war machine into a scalpel.’ (Shachtman, 2007; Cebrowski and Gartska, 1998)
Network Enabled Capability Network Enabled Capability (NEC) is a type of command and control. Command and control is the generic label for the management infrastructure behind any large, complex, dynamic resource system (Harris and White, 1987). Like all good management infrastructures, military command and control is contingent (e.g., Mintzberg, 1979) upon the problem it is tasked with dealing and because that problem, at some fundamental level, has remained relatively stable over a long period of time the term command and control has become a synonym. It has come to mean exactly, or very nearly the same thing as traditional, hierarchical, bureaucratic, centralised, ‘classic’ command and control. This is the type of command and control conjured up by images of mass battle scenes in the film Alexander, the type afflicted by the Kafkaesque pathologies inherent in all bureaucratising organisations and parodied by everyone from Black Adder to Dilbert, and the type of command and control that management books try to encourage their readers to ‘break free from’ (e.g. Seddon, 2003). The arrival of NEC shows that not all management infrastructures need to be the same. It shows that while traditional command and control is highly adept in some situations it is less effective in others. There is no one definition of NEC. In fact even the label NEC is just one in a number of ‘net-enabled’ acronyms running from Network Centric Warfare, Network Centric Operations (both favoured in North America) to Network Centric Defence (favoured in some North European countries). Regardless of acronym, the techno-organisational vision of NEC (the preferred label in the UK) refers to: ‘…self-synchronizing forces that can work together to adapt to a changing environment, and to develop a shared view of how best to employ force and
Command and Control: The Sociotechnical Perspective effect to defeat the enemy. This vision removes traditional command hierarchies and empowers individual units to interpret the broad command intent and evolve a flexible execution strategy with their peers.’ (Ferbrache, 2005, p. 104)
The logic of this organismic approach to command and control is widely held to be based on four tenets: that a) a robustly networked force improves information sharing, that b) information sharing and collaboration enhance the quality of information and shared situational awareness, that c) shared situational awareness enables selfsynchronisation, and that d) these dramatically increase mission effectiveness (CCRP, 2009). These assumptions have been appropriated by the military domain and have gathered momentum, so it is interesting to note that they are not based on much more than a decade’s worth of direct military experience. They are based on Wal Mart: ‘Here was a sprawling, bureaucratic monster of an organisation—sound familiar?—that still managed to automatically order a new light bulb every time it sold one. Warehouses were networked, but so were individual cash registers. So were the guys who sold Wal-Mart the bulbs. If that company could wire everyone together and become more efficient, then [military] forces could, too. “Nations make war the same way they make wealth”.’ (Shachtman, 2007; Cebrowski and Gartska, 1998)
This characterisation is perhaps a little crude, and perhaps not entirely fair, but it nevertheless helps us to make a powerful point that goes right to the heart of this book. Or rather, it enables Cebrowski and Gartska (1998) in their pioneering paper on the origins and future of NEC to make it for us. They say: ‘We may be special people in the armed forces, but we are not a special case. It would be false pride that would keep us from learning from others. The future is bright and compelling, but we must still choose the path to it. Change is inevitable. We can choose to lead it, or be victims of it. As B. H. Liddell Hart said, “The only thing harder than getting a new idea into the military mind is getting an old one out”.’ (Cebrowski and Gartska, 1998, p. 8)
Command and control is an organisation just like any other, conceptually at least. What has happened to make NEC the contingent response? What is it about the current state of affairs that makes NEC’s assumptions and tenets make sense? Wal Mart may be the proximal inspiration for at least some of the early thinkers on the topic of NEC but it is probably closer to the truth to say that it derives from a much wider paradigm shift, of which Wal Mart’s networked, vertically integrated operations are merely a part. Cebrowski and Gartska looked over what they saw as the artificial and contentious divide between the military arena and the rest of the world and saw that: The underlying economics have changed. ‘The new dynamics of competition are based on increasing returns on investment, competition within and between
Introduction
ecosystems, and competition based on time. Information technology (IT) is central to each of these.’ Economies are ‘characterised by extraordinary growth and wealth generation, increasing returns on investment, the absence of market share equilibrium, and the emergence of mechanisms for product lock-in. […] Locking-out competition and locking-in success can occur quickly, even overnight. We seek an analogous effect in warfare.’ (p. 1–2) The underlying technologies have changed. ‘Information technology is undergoing a fundamental shift from platform-centric computing to networkcentric computing. Platform-centric computing emerged with the widespread proliferation of personal computers in business and in the home. […] These technologies, combined with high-volume, high-speed data access (enabled by the low-cost laser) and technologies for high-speed data networking (hubs and routers) have led to the emergence of network-centric computing. Information “content” now can be created, distributed, and easily exploited across the extremely heterogeneous global computing environment.’ (p. 2–3) The business environment has changed. ‘First, many firms have shifted their focus to the much larger, adaptive, learning ecosystems in which they operate. Not all actors in an ecosystem are enemies (competitors); some can have symbiotic relationships with each other. For such closely coupled relationships, the sharing of information can lead to superior results. Second, time has increased in importance. Agile firms use superior awareness to gain a competitive advantage and compress timelines linking suppliers and customers. […] Dominant competitors across a broad range of areas have made the shift to network-centric operations—and have translated information superiority into significant competitive advantage.’ (p. 3–4)
Where Did the Humans Go? The vision just described is that of Alvin Toffler’s ‘future shock’ (1981) and ‘third wave’ (1980). While it may have been appealing and relevent there were problems with the assumptions being made about information sharing, shared situational awareness, self-synchronisation and mission effectiveness. One could look over the military/civilian divide and see these fundamental shifts not as new form of business ecology at all, but as a new form of ultra-Taylorism, a situation in which net centric technology was being pressed into the service of even greater bureaucratisation. In a lot of cases what the network seemed to be enabling were the latest Japanese business practices such as lean manufacturing, total quality management, just-in-time stock control, and all the other tools and techniques which in a dynamic and changeable commercial world enable companies, like Wal Mart, to keep at least some aspects of their complex operations in a tightly coupled and predictable state. Naturally, whilst this generally held for the benefactors of
Command and Control: The Sociotechnical Perspective
the techniques like Toyota and Wal Mart, the same could not always be said for the people at work on the production line of product or service delivery, or indeed the users of roads now overloaded with ‘just in time’ delivery traffic. The ‘technical’ aspects had clearly been optimised but at the expense of the ‘social’. Whilst from some angles these commercial systems looked like an ecosystem of self-organising flexibility, the alternative perspective was that the network had actually enabled a far closer coupled system, the ultimate high speed hierarchy. The original thinking behind NEC bears this hallmark. It started off with ‘short, decisive battles against another regular army’ in mind (Shachtman, 2007, p. 5), ‘the Soviets, the Chinese, Saddam’s Republican Guard, whoever – as long as they had tanks to destroy, territory to seize, and leaders to kill.’ (p. 5). When Cebrowski and Gartska discussed Wal Mart in the same breath as warfare what they were really driving at was: ‘a single, network-enabled process: killing’ (Shachtman, 2007, p. 8). It is no coincidence that the phrase ‘shock and awe’ (e.g., Ullman and Wade, 1996) made its way into the military lexicon at roughly the same time. But then something started to happen. After the technically optimized military successes had occurred and the mission accomplished banners had been taken down from the bridge of the USS Abraham Lincoln, a new type of organisational pathology emerged, one for which the military context seemed particularly susceptible: ‘He clicks again, and the middle screen switches to a 3-D map of an Iraqi town from a driver’s point of view. “Now let’s plan the route. You’ve got a mosque here. An IED happened over there two weeks ago. Here’s the one that happened yesterday. Hey, that’s too close. Let’s change my route. Change the whole damn thing.” He guides me through capability after capability of the command post— all kinds of charts, overlays, and animations. “But wait—there’s more,” he says. “You wanna see where all the Internet cafés are in Baghdad?” […] It’s hard not to get caught up in [the] enthusiasm. But back in the US, John Nagl, one of the authors of the Army’s new counterinsurgency manual, isn’t impressed. […] he’s more interested in what the screens don’t show. Historical sigacts don’t actually tell you where the next one’s going to be. Or who’s going to do it. Or who’s joining them. Or why. “The police captain playing both sides, the sheikh skimming money from a construction project, […] what color [icon] are they?”.’ (Shachtman, 2007, p. 5)
The paradox in all this is that if, as Cebrowski and Gartska acknowledge, ‘military operations are enormously complex, and complexity theory tells us that such enterprises organise best from the bottom-up’ (1998, p. 4–5) then what happened to the most important low-level component of them all: the human? The focus on Toffler’s third wave of fundamental changes in economics, technology and business has tended to put the focus on the ‘network’ rather than what it ‘enables’. Which is for the most agile, self-synchronous component of all in NEC,
Introduction
the people, to use its capability to perform the one task they excel at: coping with complexity. For the human factors practitioner encountering the extant literature on NEC for the first time the overriding feeling is undoubtedly one of opportunity. Here we have a huge practical domain of interdisciplinary science where, for once, the human is widely acknowledged to be key. Within this expanding literature are concepts and ideas which have tremendous potential, not only for existing forms of human factors but for fundamentally new types of human factors. Yet for all the opportunity one has to acknowledge some occasional disappointment. In a very real sense, where did the human go? Speaking from a human factors perspective, some of the more military orientated work in this domain seems stronger on doctrine and box diagrams than it does on theory and evidence. Particularly noteworthy are concepts with a long legacy and substantial theoretical underpinning in human factors, re-invented without a great deal of reference to what has gone before, or worse, the human in NEC is reduced to a form of rational optimiser and mathematically modelled from there. If that is the human factors view, then from the other side of the fence the military reader looking in would probably be inspired and disappointed in equal measure. Common criticisms of human factor’s initial forays into NEC are often rooted in what is perceived as a lack of foundational rigour and mathematical profundity, something that can, and routinely is, levelled at all social sciences. This combines with what can often seem like naivety and a lack of appreciation for military context. All told, the human in NEC represents a very difficult interface and it is clear that NEC is uncommonly prone to scientific antagonism of precisely this sort. Quite simply, every time one crosses an interdisciplinary barrier, of which there are many in NEC research, there is the ever present risk of ruffling feathers and seeming to ride roughshod over any number of finely tuned concepts. It would certainly be easy for the work presented in this volume to be construed in exactly the same way, but that would be to mistake its message. If the nexus of various disciplines impinging on NEC are seen as a form of Venn diagram then we are not necessarily concerned with the larger parts of allied disciplines which do not overlap and for which a human factors approach is not appropriate; instead we are concerned with the smaller areas which do overlap. This book, rather like the fundamental systems concepts underlying NEC itself, is not focused merely on parts but on the connections between those parts. In line with this motif we seek to build linkages, however tentative, that extend the reach of human factors towards other domains such as organisational theory, complex systems research and military operations. For the reader approaching the work from any of these individual specialisms, there may not be much that is fundamentally new apart from the way it has been applied. Rather than re-invent the wheel and engage in an antagonistic form of science, quite the opposite is intended. The spirit in which this book is written is for these linkages to represent bridgeheads from which human factors tries to reach out across its various interdisciplinary boundaries in the hope that other disciplines can see enough that is relevant to
Command and Control: The Sociotechnical Perspective
want to reach over from the other side. The linkages may at times appear tentative, focusing on breadth rather than depth, incomplete, crude even, but no apology is made for trying to establish what constitutes a stating point. It is from such a point that those within the NEC research community, human factors included, can begin to speak to each other. The message is one of reconciliation, of an interconnected scientific approach that could potentially become, like NEC itself aspires, to become more than the sum of its disciplines. Sociotechnical Theory John Gartska was interviewed by a journalist in 2007 and rightly defends the NEC concepts he helped to set in motion with his and Cebrowski’s paper. Even he, though, acknowledged that in the short space of a decade the net-centric vision has changed: ‘You have to think differently about people’ he was noted as saying. ‘You have your social networks and technological networks. You need to have both.’ (Shachtman, 2007, p. 8)
This is the precise sentiment behind the work contained in this volume. In human factors literature (and beyond) the word sociotechnical is ubiquitous. On the one hand you have the socio, of people and society, and on the other the technical, of machines and technology. Socio and technical combine to give ‘sociotechnical’ (all one word) or ‘socio-technical’ (with a hyphen). Both variations mean the same thing. In the human factors world we speak freely of ‘purposeful interacting socio-technical systems…’ (Wilson, 2000, p. 557), ‘complex sociotechnical systems …’ (Woo and Vicente, 2003, p. 253), ‘sociotechnical work systems…’ (Waterson, Older Gray and Clegg, 2002, p. 376) and many more besides. But what does it actually mean? Like command and control, the phrase ‘sociotechnical’ has become a synonym, a descriptive label for any practical instantiation of socio and technical, people and technology, the soft sciences meeting hard engineering. In these terms sociotechnical does not mean a great deal as it is difficult to imagine any meaningful system these days that is not described by these two worlds colliding. Whether designed, manufactured, used, maintained or disposed by humans, humans are involved somewhere along the line and the term sociotechnical emerges merely as a convenient, if slightly tautological, buzzword for what results. There is more to the phrase sociotechnical than this, however. In fact, sociotechnical has a very specific meaning. Sociotechnical ‘theory’ refers to the interrelatedness of ‘social’ and ‘technical’ and is founded on two main principles. One is that the interaction of social and technical factors creates the conditions for successful (or unsuccessful) system performance. These interactions are comprised partly of linear cause and effect
Introduction
relationships, the relationships that are normally designed, and partly from nonlinear, complex, even completely unpredictable and chaotic relationships, those that are unexpected. Whereas the ‘technical’ is assumed to be the former (i.e., linear), the latter (i.e., non-linear) tends to be the latter domain of the ‘socio’; after all, people are not rational optimisers and they do not behave like machines. Then again, as machines and technology venture further along the transformation from platform centric to net-centric, when even simple-minded technology is connected together, what often results is the unpredictable, emergent, non-linear behaviours more associated with the socio than the technical. Inevitably, then, what one usually gets when socio and technical systems are put to work are both types of interaction: linear and non-linear, safe and predictable versus unsafe and unpredictable, effects that are directly proportionate to the sum of their causes combined with other effects that are grossly disproportionate. This is the first principle. Sociotechnical theory’s answer is the second of its two main principles: joint optimisation. The argument is that the optimisation of either socio, or far more commonly, the technical, tends to increase not only the quantity of unpredictable, un-designed, non-linear relationships but specifically those relationships that are injurious to the system’s performance. Here we are talking of self-destruction rather than self-synchronisation. How does one achieve joint optimisation? The further argument put forward by the sociotechnical school of thought is to combine the conceptual goal of sociotechnical theory with systems theory to give ‘sociotechnical systems theory’ (SST). The term sociotechnical ‘system’ is also frequently used as a descriptive label, as a synonym for any mixture of people and technology, but like sociotechnical ‘theory’ it refers to an explicit set of concepts deployed in order to design organisations that exhibit open systems properties, and by virtue of this, cope better with environmental complexity, dynamism, new technology and competition. The main open systems metaphors developed and deployed by SST are described in Chapter 2. Parallel Universes Chapter 2 demonstrates that the parallel universes of NEC and sociotechnical systems theory overlap. This is the first contention put forward by this book; that NEC is not alone. Where NEC speaks of peer-to-peer interaction and adhoc groups, SST speaks of semi-autonomous groups. Where the former speaks of effects based operations the latter speaks of minimal critical specification, and so it goes on, to such an extent that there is considerable and almost complete overlap. The key difference is that whereas the former is based on around 10 years of military application (with promising but undeniably mixed results) the latter is based on over 50 years of commercial application with an overwhelming tack
Command and Control: The Sociotechnical Perspective
record of success. It would indeed be ‘false pride’ to not explore such a symmetry further. Chapter 3 continues to work on this contention. Contemporaneous with SST’s birth in the 1950’s is the overlapping theme of social network analysis. NEC is overlain across a classic empirical study into the effects of different network structures, not just in terms of their outright performance but also the experience of those at work within them. Joint optimisation in the classic sociotechnical sense sees the structure of an organisation as a major tool in helping to create the conditions for cohesion, trust, shared awareness and all the other facets of joint optimisation and self-synchronisation. This link is made explicit in Chapter 3 and developed throughout the book. Matching Approaches to Problems Chapter 4 interrupts the progression from SST’s considerable legacy in the commercial world and the supporting empirical studies contained in Chapter 3 by getting in touch with a concept fundamental to both SST and NEC, that of complexity. The question is directed to exactly what NEC and SST are an organisational response to? Taking a lead from the extant NEC literature the answer is clearly complexity. The difference here is that the concept is considered from a social/human science perspective. Despite increasing lip service being paid to it in human factors literature, the study of complexity has not generally been followed up with an in-depth exploration of what it actually is and what it means. Chapter 4 seeks to address this and align human factors with NEC in this important respect. It does so by pulling through a number of key concepts from complex systems research and recasting them in a human factors mould. In the course of doing this we put forward a number of arguments to support our second contention, that if complexity is such a good word for describing many human factors problems (not least NEC) then the corresponding human factors approach has to be matched to it in quite a fundamental way. What is the Network Enabling? Chapter 5 continues the progression from legacy to live-NEC. Consistent with matching approaches to problems the NATO SAS-050 approach space, a popular and widely accepted model of command and control, is extended through the use of social network analysis and applied in a live context. Reference to social network analysis has already been made in Chapter 3, and it combines with the insights presented in Chapter 4 to inspire the method by which we have been able to turn the NATO SAS-050 model of command and control from a typology into a practical, usable taxonomy.
Introduction
Numerical metrics can be extracted from large scale case studies, enabling the approach space to be populated with live data and for the organisation’s structure to determine where in the approach space it resides. This answers the fundamental question as to what type of organisation NEC actually is. If structure is an important determinate of joint optimisation then it is of further interest to note that, in real life, NEC changes its structure contingent on both function and time. This links to our third contention, which is that the organisational agility and tempo observed during a live case study results in large part because of the humans in the system. Moreover, it was due in similarly large part to the ‘unexpected’ adaptations they undertook when faced with the initial conditions that the NEC system in this case represented. In other words, the actual region in the approach space adopted by the organisation did not match the expected region. We examine why. An Interconnected Approach to Analysis Now armed with live and empirical data from Chapters 3 and 5 respectively, Chapter 6 sets it once again against a complexity related backdrop. Now data driven, further insights can be pulled through and mapped onto human factors. In this context sensitive dependence on initial conditions relates to a shift away from assuming the human system interaction to be stable, to instead assuming to be inherently unstable. The nature of this fundamental instability is considered in turn from the perspective of emergence. Tentative steps are taken towards understanding this concept from a human perspective and leveraging such insights to determine what human factors methods are best suited to understanding it. Next, the link from complex systems research to the NATO SAS-050 approach space is further elaborated, as it is from this that the ‘extended’ approach space was inspired. By these means the NATO SAS-050 approach space gains a sizeable and important retrospective legacy. From Theory to Practice Chapter 7 maintains the level of analysis at that of the total system by continuing to deploy the extended NATO SAS-050 approach space. Having accumulated a body of evidence in favour of a sociotechnical approach to NEC system design, the question arises as to specifically what form should this take? What organisational structures are expressive of joint optimisation as well as representing a match in terms of approach and putative problem? In order to explore this high level question the NATO SAS-050 model is populated with a wide variety of organisations, from commercial concerns through to terrorist networks. This enables the live instance of NEC presented in Chapter 5 to be set into some kind of real-world context. From this analysis it emerges that many of the organisations which seem to cope with the challenges of the information-age posses the structural properties of
10
Command and Control: The Sociotechnical Perspective
a ‘small-world’ network. To the extent that this configuration is useful, how does one go about designing an organisation to exhibit such properties? The answer, surprisingly, is to design it according to sociotechnical principles. In this we see, once again, demonstrable evidence that the sociotechnical approach overlaps with yet another of the disciplines that impinge on NEC, not just social network analysis, complexity and organisational science, but also cutting-edge research into network topologies. The sociotechnical approach, therefore, is expressive of the book’s fourth main contention, which is to advocate (despite ruffled feathers and occasional analytic crudities) an interconnected approach to analysis and to NEC’s manifold issues, significant amongst which is of course the human. In Chapter 8 the perspective inverts and the analysis descends from the topdown view of jointly optimising network structures to the bottom-up view of the individual ‘networked interoperable components’ that reside within them. With equipment that is journeying along the transition from platform-centric to netcentric the corresponding human factors approach taken towards them requires scrutiny. In this we see evidence of NEC research informing human factors, not merely the other way around. We argue that in the face of such a transition the systems concepts which apply so well at the level of the total system now also apply to the level of the components, or at least those which increasingly exhibit more systemic behaviour. Inspired by the extent of unexpected adaptation in the real-world case of NEC covered in Chapter 5, as well as the extant literature, the final contention is that the goal of human centred design shifts from designing an end product towards the creation of entities and artefacts that encourage favourable adaptations from initial conditions. In other words, equipment needs to be designed so that it is as agile and self-synchronising as the system within which it resides. This in turn requires a new domestication of systems concepts. Summary Despite the various arguments, contentions and themes, the aim of this book is an overridingly simple one: to apply SST to the practice of NEC system design. This task is undertaken against a wider backdrop of Human Factors Integration (HFI). If HFI is about ‘…providing a balanced development of both the technical and human aspects of equipment provision’ (MoD, 2000, p. 6) then this book sets out to describe the way in which SST offers a set of theories, empirical evidence, practical measures, and most importantly, a legacy of delivering on such objectives.
Chapter 2
Reconsidering the Network in NEC Aims of the Chapter The key issue for traditional hierarchical command and control is that it is increasingly challenged by a host of distinctly modern problems, namely, environmental complexity, dynamism, new technology, and competition that is able to exploit the weaknesses of an organisational paradigm that has been dominant since the Industrial Revolution. The conceptual response to these challenges is NEC. The aim of this chapter is to show that although developed independently, NEC exhibits a high degree of overlap with concepts derived from sociotechnical systems theory (SST). Sociotechnical theory brings to NEC research much more than merely conceptual overlap. It also brings a successful 60-year legacy in the application of open-systems principles to commercial organisations. This track record is something that NEC research currently lacks. This chapter reviews the twin concepts of NEC and sociotechnical systems theory, the underlying motivation behind the adoption of open systems thinking, provides a review of classic sociotechnical studies and the legacy that comes with it. It is argued throughout that ‘classic’ sociotechnical systems theory has much to offer ‘new’ command and control paradigms. Rationality and Industrial Age Thinking ‘Formal organisation design, or deliberate as opposed to informal or evolved organisation design, is part of the evolution of both Western and Eastern civilizations’ (Davis, 1977, p. 261). Organisations are more or less synonymous with the descriptive use of the term ‘sociotechnical system’ in that they are invariably a mixture of socio and technical. Not all organisations, however, are sociotechnical systems in the systemic, jointly optimised usage of the term (as is used in this book). To be clear, organisations, of which military command and control is but one example, are entities to which sociotechnical systems theory can be applied. In virtually all developed civilisations the recent history in organisational design is wedded to a shared ‘industrial age’ mindset (Beringer, 1986), one that forms the backdrop to both sociotechnical systems and NEC, and one that can be explained with reference to a deeper, more fundamental concept. Formal rationality (Weber, 2007; Ritzer, 1993; Trist, 1978) lends weight to the opinion of many eminent observers over the years who have been exercised not by the apparent mastery of
Command and Control: The Sociotechnical Perspective
12
human endeavour over nature but instead the various maladies. None other than Elton Mayo (1949) surveyed the state of affairs yielded by 100 years of industrial activity and wrote, ‘To the artist’s eye, some-thing was decidedly eschew in the actual Victorian progress; and that something continues to this day’ (p. 4). In spite of all the time, effort and expense that feeds into the design of organisations (e.g., Ritzer, 1993; Davis, 1977), systems (e.g., Bar-Yam, 2003), major projects (e.g., Morris and Hough, 1987) consumer products (e.g., Green and Jordan, 1999) and now NEC (e.g., Baxter, 2005), what consistently emerges is something that is often substantially less effective than intended (Clegg, 2000). More than that, these are entities and artefacts that are occasionally injurious to human well being (although technically effective they are often criticized for being ‘dehumanizing’; Ritzer, 1993) and may, in extreme cases, become ‘anti-human’. In military command and control the organisational aetiology of friendly fire incidents seems to be a case in point. Formal rationality is a prominent part of an ‘implicit theory’ that has guided modern organisational design since the Industrial Revolution. A formally rational organisation is labelled a bureaucracy, in the scientific rather than pejorative sense, and the stereotypical case of so-called ‘classic’ hierarchical command and control (C2) fits into this category well. Rationalising organisations exhibit a tendency towards hierarchies, reductionism and the maximisation of the following attributes: • • •
•
Efficiency: A rational organisation is ‘…the most efficient structure for handling large numbers of tasks…no other structure could handle the massive quantity of work as efficiently’ (Ritzer, 1993, p. 20), Predictability: ‘Outsiders who receive the services the bureaucracies dispense know with a high degree of confidence what they will receive and when they will receive it’ (Ritzer, 1993, p. 21), Quantification: ‘The performance of the incumbents of positions within the bureaucracies is reduced to a series of quantifiable tasks…handling less than that number is unsatisfactory; handling more is viewed as excellence’ (Ritzer, 1993, p. 21), Control: ‘…the bureaucracy itself may be seen as one huge nonhuman technology. Its nearly automatic functioning may be seen as an effort to replace human judgement with the dictates of rules, regulations and structures’ (Ritzer, 1993, p. 21).
Like all bureaucracies, classic C2 rests on ‘tried and true assumptions: that the whole will be equal to the sum of the parts; that the outputs will be proportionate to the inputs; that the results will be the same from one application to the next; and most fundamentally, that there is a repeatable, predictable chain of causes and effects.’ (Smith, 2006, p. 40). As a result, one metaphor for a bureaucracy is as a type of ‘organisational machine’ (Arnold, Cooper and Robertson, 1995). In other words, ‘When all the incumbents have, in order, handled the required task, the
Reconsidering the Network in NEC
13
goal is attained. Furthermore, in handling the task in this way, the bureaucracy has utilised what its past history has shown to be the optimum means to the desired end’ (Ritzer, 1993, p. 20). This is the nub of what formal rationality is really all about. In summary, then, organisations designed along bureaucratic lines can be seen as a way of imposing control theoretic behaviour on a large scale, and in so doing, trying to make inputs, processes, outputs, and even humans, behave deterministically. The core tenets of formal rationality – efficiency, predictability, quantification and control – in turn link to a much more recent model of command and control developed by NATO working group SAS-050 (NATO, 2007). This model provides three major axes (and a three dimensional space) within which various instantiations of command and control can be plotted. The purpose of defining this so-called ‘approach space’ is to explore alternative paradigms for command and control, ones that are becoming increasingly tractable with the growth in networked technologies. The formally rational instance of classic, hierarchical C2 can be positioned in the NATO SAS-050 model as shown in Figure 2.1. This type of organisation might be characterised by unitary decision rights (in which
NEC
Patterns of Interaction
Distributed
Hierarchical
100
50
0 100
n of Dec ision Rights Pee
Figure 2.1
eer r-to-P
0
-50
n of ibutio Distr rmation o f In
-100 ad
Allocatio
Tight
Classic C2
50
Bro
0 ry Unita
The NATO SAS-050 approach space positions traditional hierarchical command and control in a three-dimensional space defined by unitary decision rights, hierarchical patterns of interaction and tight control of information
14
Command and Control: The Sociotechnical Perspective
optimum means to ends are specified at the top of, or at higher levels of, a vertical hierarchy); tightly constrained patterns of interaction (owing to rules, standard operating procedures and other means by which the bureaucracy embodies its past experience and specifies optimum means to ends) and tight control (in which performance can be quantified and controlled through intermediate echelons of management). It is these features, and their formally rationalistic backdrop, that together make up the diffuse zeitgeist referred to in contemporary NEC literature as ‘industrial age thinking’ (e.g., Smith, 2006; Alberts, 2003; Alberts, Garstka and Stein, 1999; Alberts and Hayes, 2006; Alberts, 1996). The Irrationality of Rationality The tension created by this prevailing climate of industrial age thinking emerges from ‘…the observation that Rational systems, contrary to their promise, often end up being quite inefficient’ (Ritzer, 1993, p. 122). As Ritzer (1993) goes on to explain: ‘Instead of remaining efficient, bureaucracies can degenerate into inefficiency as a result of “red tape” and the other pathologies we usually associate with them. Bureaucracies often become unpredictable as employees grow unclear about what they are supposed to do and clients do not get the services they expect. The emphasis on quantification often leads to large amounts of poor-quality work…All in all, what were designed to be highly Rational operations often end up growing quite irrational’ (Ritzer, 1993, p. 22). Experience over centuries of conflict (e.g., Regan, 1991) make it possible to go further to say that in some cases classic C2 can actively create inefficiency (instead of efficiency), unpredictability (instead of predictability), incalculability (instead of calculability) and a complete loss of control (Ritzer, 1993; Trist and Bamforth, 1951). These are the antithetical problems, ironies and productivity paradoxes that, when all else fails, tend to fall into the lap of human factors. The overarching source of these problems, simply stated, is that despite attempts to impose deterministic behaviour on an organisation and its environment, the resultant interlinked entities, or ‘system’, is actually very hard to maintain in a fixed state (or as something analogous to a ‘closed system’). All the more so when such systems are subject to a large range of external disturbances. In fact, the greater the scale and extent of determinism that is imposed the worse the problem can actually become. As these systems grow larger and more interlinked the bigger the effect that such organisations have on their environment. They end up becoming the environment and a prime cause of non-linear change and complexity within it (Emery and Trist, 1965). The evolution of military command and control is particularly instructive. The apotheosis of classic C2 was seemingly reached in the large scale techno-centric style of attrition seen in the cold war era (Smith, 2006). Technically effective hardly seems an adequate term for the sheer destructive might of the opposing military forces in question. Their evolution, for many lesser organisations with militaristic ambitions, served to fundamentally change the environment within
Reconsidering the Network in NEC
15
which such activities took place. A new paradigm emerged: asymmetric warfare. Asymmetric warfare is characterised by largely urban operations, an opposing force that co-exists with a civilian populous, one that does not adhere to ‘rules of engagement’ of the sort that classic C2 is adapted. Ironically, the type of organisation that classic C2 now faces is altogether more swarm-like, agile and self-synchronising. It is an organisation that exhibits open systems properties to a far greater degree than the highly technically effective organisation to which these activities are often directed. Is the current situation a case of closed versus open systems? Perhaps. Either way, military organisations around the world, to paraphrase the sociotechnical pioneers Trist and Bamforth (1951), ‘seem ready to question a method which they have previously taken for granted’. Network Enabled Capability Sitter, Hertog and Dankbaar (1997) offer two solutions for organisations confronted with such difficulties. With an environment of increased (and increasing) complexity, ‘The first option is to restore the fit with the external complexity by an increasing internal complexity. […] This usually means the creation of more staff functions or the enlargement of staff-functions and/or the investment in vertical information systems’ (p. 498). And the second option? ‘…the organisation tries to deal with the external complexity by “reducing” the internal control and coordination needs. […] This option might be called the strategy of “simple organisations and complex jobs”.’ This provides a neat characterisation for the current state of affairs in closed versus open systems or asymmetric warfare. One way of putting it is that a traditional military organisation, a complex organisation with simple jobs, is facing a simple organisation with complex jobs. The purist vision of NEC is congruent with the latter option. Referring back to the NATO SAS-050 model of command and control presented earlier, it can be seen that NEC, unlike classic C2, is characterised by broad dissemination of information and shared awareness. This ‘includes the sharing not only of information on the immediate environment and intentions of our own enemy and neutral forces, but also the development of shared combat intent and understanding’ (Ferbrache, 2005, p. 104). NEC is further characterised by distributed patterns of interaction and agility. This is the ‘ability to reconfigure forces and structures rapidly, building on this shared awareness, exploiting effective mission planning methods, and enabled by an information environment that allows rapid reconfiguration of the underlying network and knowledge bases’ (Ferbrache, 2005, p. 104). NEC is also characterised by peer-to-peer interaction and synchronisation, which is the ‘ability to plan for and execute a campaign in which we can ensure all elements of the force work together to maximum military effect by synchronizing the execution of their missions to mass forces or generate coordinated effects on target’ (Ferbrache, 2005, p. 104). Although there is little evidence of overt cross-referencing, this vision is shared almost exactly with the ‘simple organisations and complex jobs’
Command and Control: The Sociotechnical Perspective
16
concept, in other words, those organisations designed according to sociotechnical principles. In the remainder of the chapter our primary task will be to explore the domain of sociotechnical systems theory (SST) and to review exactly how it could bestow NEC-like open systems behaviour upon military command and control. This task will be conducted against a backdrop of continually trying to establish synergies between these two seemingly compatible domains. Main Principles of Sociotechnical Theory Sociotechnical theory offers a theoretical basis from which to design organisations, and moreover, to harness the advantages that NEC-like command and control promises. This section goes back to first principles and the seminal work of Trist and Bamforth (1951) entitled, ‘Some social and psychological consequences of the longwall method of coal getting’. Despite the coal mining subject matter this is a fascinating and surprisingly relevant case study. The analysis was motivated by the following irrationality: ‘Faced with low productivity despite improved equipment, and with drift from the pits despite both higher wages and better amenities […] a point seems to have been reached where the [coal] industry is in a mood to question a method it has taken for granted’ (Trist and Bamforth, 1951, p. 5). The longwall method of coal mining represented something akin to mass production. Large-scale machinery was able to cut large swathes of material from the face and represented a considerable economy in man power and the potential for high productivity. At a more fundamental level this method of working also reflected a number of rationalistic principles: • • •
Large-scale coal cutting machinery led to a simplification and specialisation of the miners tasks. The method of working became driven by the needs of the mechanised method with the pattern of the shift and its social organisation changing as a result. This new organisation required an intermediate level of supervision and management that was previously absent.
The NATO SAS-050 approach space would characterise the longwall method of coal getting by its hierarchical patterns of interaction (the large scale operations required the task to be decomposed…), its unitary allocation of decision rights (…with one person in charge of the shift…) and its relatively tight control over the distribution of information (…due to increasing role specialisation). All of this means that it is conceptually very similar to classic C2 despite the vast differences in task type and domain. In sociotechnical terms, the new mechanised longwall
Reconsidering the Network in NEC
17
method of mining represented something of a retrograde step compared to the previous ‘hand got method’. The hand got method divided up the coal getting task along the coal face, with miners working in closely knit teams, mining a certain section of the face from start to finish with all members of the team able to undertake all parts of the task. This method, despite its arcane outward appearance, was nevertheless characterised by a broader dissemination of information (all team members knew what was expected of them…), more distributed patterns of interaction (…the structure of the shift was less hierarchical) and more devolved decision rights (…with greater autonomy bestowed upon the team). Trist and Bamforth explain that, ‘the longwall method [can be] regarded as a technological system expressive of the prevailing outlook of mass-production engineering and as a social structure consisting of the occupational roles that have been institutionalized in its use’ (1951, p, 5). The prevailing outlook that they refer to is the industrial age rationalizing method of the ‘factory system’ (the Fordist production line) applied to coal mining. They continue: ‘These interactive technological and sociological patterns will be assumed to exist as forces having psychological effects in the life-space of the face worker, who must either take a role and perform a task in the system they compose or abandon his attempt to work at the coal face’ (p. 5). The psychological effects of the interacting socio and technical ‘forces’ (to use the language of 1950s sociotechnical theory) was leading to reduced productivity and increased absenteeism. The notion of ‘interactive technological and sociological patterns’ quickly evolved to become the term ‘sociotechnical’. In order to explore this productivity paradox 20 miners with varied experience of the work domain were interviewed at length, along with various management and training role incumbents. Trist and Bamforth’s paper is thus based on ethnographic techniques, the outcomes of which led to the elaboration of a number of enduring sociotechnical principles which include the following: Responsible Autonomy ‘The outstanding feature of the social pattern with which the pre-mechanized equilibrium was associated is its emphasis on small group organisation at the coal face’. Indeed, ‘under these conditions there is no possibility of continuous supervision, in the factory sense, from any individual external to the primary work group’ (Trist and Bamforth, 1951, p. 7). Harsh physical constraints quite simply prevented the task from being carried out, with this technology, in any other way. So instead of a larger ‘whole of shift’ based organisation accountable to intermediate layers of management, the hand-got method embodied internal supervision and leadership at the level of the ‘group’ which resulted in so-called ‘responsible autonomy’ (Trist and Bamforth, 1951, p. 6). Sociotechnical theory was pioneering for its focus on the group as the primary unit of analysis.
18
Command and Control: The Sociotechnical Perspective
A facet of this method of working that is somewhat unique to the dangers of the underground situation, yet with ready parallels to military operations, is ‘the strong need in the underground worker for a role in a small primary group’ (Trist and Bamforth, 1951, p. 7). It is argued that such a need arises in hazardous circumstances, especially in cases where the means for effective communication are limited. As Carvalho (2006) states, operators use this proximity and group membership ‘…to produce continuous, redundant and recursive interactions to successfully construct and maintain individual and mutual awareness…’ (p. 51). The immediacy and proximity of trusted team members makes it possible for this need to be met with favourable consequences for team cohesion, trust and overall system performance. The field of NEC is rightly interested in the varied issues of trust and team cohesion (e.g., Siebold, 2000; Oliver, et al., 2000; Mael and Alderks, 2002) and whilst the principles of sociotechnical systems theory are far from a panacea they do at least admit the possibility of creating favourable conditions for these varied aspects to emerge. Adaptability As Trist and Bamforth put it, ‘though his equipment was simple, his tasks were multiple’, the miner ‘…had craft pride and artisan independence’ (1951, p. 6). The ‘hand-got method’ is an example of a simple organisation (and equipment) ‘doing’ complex tasks. The longwall method, on the other hand, is an example of a complex organisation (and technological infrastructure) ‘doing’ simple tasks. Job simplification has long been associated with lower moral and diminished job satisfaction (e.g., Hackman and Oldman, 1980; Arnold, Cooper and Robertson, 1995). In the former case a type of ‘human redundancy’ was created (e.g., Clark, 2005) in which ‘groups of this kind were free to set their own targets, so that aspiration levels with respect to production could be adjusted to the age and stamina of the individuals concerned’ (Trist and Bamforth, 1951, p. 7). This meant that outcomes or ‘effects’ were more important than activities or the precise means by which those effects were achieved. Trist and Bamforth (1951) go on to note that ‘A very large variety of unfavourable and changing environmental conditions is encountered at the coal-face, many of which are impossible to predict. Others, though predictable, are impossible to alter.’ (p. 20). The longwall method was clearly inspired by the appealing industrial age, rational principles of ‘factory production’ wherein ‘a comparatively high degree of control can be exercised over the complex and moving ‘figure’ of a production sequence, since it is possible to maintain the ‘ground’ in a comparatively passive and constant state’ (Trist and Bamforth, 1951, p. 20). In many contexts, coal mining and military operations being just two, there is relatively little in the way of opportunity for maintaining the ‘ground’ in such a state. If the environment cannot be made to approximate to the linear cause and effect logic of determinism
Reconsidering the Network in NEC
19
then ‘the applicability […] of methods derived from the factory’ is limited (Trist and Bamforth, 1951, p. 20). Meaningfulness of Tasks Sociotechnical theory is as concerned for the experience of humans within systems as it is with the system’s ultimate performance. Under the rubric of joint optimisation sociotechnical systems theory sees the two as complementary. Trist and Bamforth (1951) go into detail as to how this complementary relationship was realised in their mining example. They identify the hand-got method as having ‘the advantage of placing responsibility for the complete coal-getting task squarely on the shoulders of a single, small, face-to-face group which experiences the entire cycle of operations within the compass of its membership.’ Furthermore, ‘for each participant the task has total significance and dynamic closure’ (Trist and Bamforth, 1951, p. 6). It is a meaningful task. Meaningfulness arises out of a focus on the group, from responsible autonomy and from adaptability, linking jointly optimised sociotechnical systems to a number of ‘core job characteristics’ (Hackman and Oldman, 1980): • • • • •
Skill variety (e.g., simple organisations but complex varied jobs that rely on a multiplicity of skills; Sitter et al., 1997). Task Identity (e.g., ‘entire cycle of operations’ or whole tasks; Trist and Bamforth, 1951). Task Significance (e.g., ‘dynamic closure’ and meaningful tasks; Trist and Bamforth, 1951). Autonomy (e.g., human redundancy, adaptability, semi-autonomous work groups; Trist and Bamforth, 1951). Feedback (e.g., continuous, redundant and recursive interactions; Carvalho, 2006).
The pioneering work of Trist and Bamforth was motivated by the most prominent irrationality of rationality, namely, that organisations designed to function deterministically like machines are dehumanising. It elaborated on the central themes that would form a fully fledged sociotechnical school of thought, as well as the specific features that would bestow desirable open systems characteristics on organisations which previously behaved in a way analogous to closed systems. There are ready parallels between Trist and Bamforth’s foundational work in the Durham coal mines and the dynamic, uncertain and often dangerous world of military operations. Sociotechnical-systems theory provides a detailed conceptual language with direct links to NEC system design. Perhaps the major influence of Trist and Bamforth’s pioneering sociotechnical work, however, was to change the prevailing viewpoint in which organisations were considered: from a purely technical perspective (industrial age thinking) or as purely social entities (an
20
Command and Control: The Sociotechnical Perspective
organisational or industrial relations perspective) to instead ‘…relate the social and technological systems together’ (Trist, 1978, p. 43). Sociotechnical Systems Systems thinking offered the conceptual language from which notions of ‘networks’ and ‘distributed systems’, and indeed, NEC itself ultimately derive. Its application to sociotechnical theory came in 1959 with a paper by Emery which expanded the field by drawing on the specific case of open systems theory (Kelly, 1978, p. 1075; Bertalanffy, 1950). Open systems theory gave sociotechnical theory a more tightly defined grounding as well as a unifying conceptual language. The characteristics of systems thinking are the twin notions of ‘a complex whole’ formed from a ‘set of connected things or parts’ (Allen, 1984). Part of the appeal of industrial age thinking is that the ‘set of connected things or parts’ can be tightly defined. A visual metaphor for such a deterministic system might be an electrical circuit diagram, an artefact with components that have known input/output properties connected by electrical pathways with similarly known properties and flows. Such an artefact, in systems terms, could be referred to as a closed system or an ‘object’ or a rational system. We of course have to rely on a degree of analogy here. Whilst not seeking to push the analogy too far, we can state that in organisational terms a closed system can be said to be concerned with the attainment of a specific goal and there is a high degree of formalisation (Scott, 1992). Another word used to describe such an organisation is a bureaucracy. An open system is quite different. Import and Export Open systems are acknowledged to have boundaries with other systems and some form of exchange existing between them: ‘A system is closed if no material enters or leaves it; it is open if there is import and export and, therefore, change of the components’ (Bertalanffy, 1950, p. 23). In the original biological conception of open systems this exchange would be ‘matter’ such as haemoglobin or oxygen. As systems theory has expanded, the inviolable characteristic of all such exchanges is now seen as essentially informational (e.g., Kelly, 1994; Ciborra, Migliarese and Romano, 1984). Exchange between system elements is input, which causes state changes, outputs of which become further inputs for other elements. An appropriate visual metaphor for such a system might be a block or Venn diagram in which the properties of the components and the links between them are not as well defined as an electrical circuit diagram and are subject to change. A system exhibiting these properties is also referred to as a network, expanding somewhat the definition of ‘network’ in NEC from that of a computer network to a formal systems metaphor.
Reconsidering the Network in NEC
21
Steady States ‘A closed system must, according to the second law of thermodynamics, eventually attain a time-independent equilibrium state, with maximum entropy and minimum free energy […] An open system may attain (certain conditions presupposed) a time-independent state where the system remains constant as a whole… though there is a constant flow of the component materials. This is called a steady state’ (Bertalanffy, 1950, p. 23). If sociotechnical systems are open systems then organisations become analogous to a ‘vitalistic’ (i.e., living) entity. The idea of a ‘vitalistic entity’ strikes a chord in organisational psychology. At least one metaphor for an organisation is ‘organismic’, meaning that it is able to adapt and evolve to environmental changes and behave more like an ecology than a machine (e.g., Morgan, 1986). This is something that NEC, with its terminology couched very much in terms of ‘agility’ and ‘self-synchronization’ is undoubtedly, if sometimes implicitly, striving for. By comparison, a closed system is, or becomes, locked or frozen in a particular state and requires no further import or export of information to maintain that state. A closed system (or industrial age organisation taken to its extreme) is therefore unresponsive to environmental change, matched to an optimum means to an end within a defined context and slow to change or adapt, if at all. To use a computer science metaphor, a closed system could be described as an entity that is ‘programmed’ while an open system is an entity that can ‘learn’. Equifinality Related to these ideas of dynamism and adaptability is the notion of equifinality, described by Von Bertalanffy thus: ‘A profound difference between most inanimate and living systems can be expressed by the concept of equifinality. In most physical systems, the final state is determined by the initial conditions… Vital phenomena show a different behaviour. Here, to a wide extent, the final state may be reached from different initial conditions and in different ways’ (p. 25). This is exactly what NEC desires when it speaks of self-synchronisation and what sociotechnical theory offers in terms of adaptability and semi-autonomy. Equifinality grants open systems a certain ‘paradoxical behaviour’, ‘as if the system “knew” of the final state which it has to attain in the future’ (p. 25), which of course it does not, merely that sociotechnical principles permit it to rapidly adapt and evolve to one. Trist (1978) could have been describing NEC when saying that open systems grow ‘by processes of internal elaboration. They manage to achieve a steady state while doing work. They achieve a quasi-stationary equilibrium in which the enterprise as a whole remains constant, with a continuous “throughput”, despite a considerable range of external changes.’ (p. 45). To sum up, then, sociotechnical systems theory is inextricably linked to ideas about open systems. The principles first elaborated by Trist and Bamforth (1951) are framed in terms of endowing favourable open
22
Command and Control: The Sociotechnical Perspective
systems behaviour on organisations. Military command and control is one such organisation. Real-Life Sociotechnical Systems Classic Examples Since the pioneering work of Trist and Bamforth in 1951 there has been considerable effort undertaken in organisational design using sociotechnical principles. During the late sixties and seventies, riding on a groundswell of left-leaning ideologies, sociotechnical systems theory was attractive. Indeed, it was the darling of business process engineering occupying a niche which today might be occupied by TQM and Six Sigma. A considerable number of organisations embraced sociotechnical systems methods, the two most famous being Volvo and Olivetti. We begin this section by providing a case study from the ‘classic’ sociotechnical school. The intention is to provide a grounding and contextualisation, describing what an organisation redesigned according to sociotechnical principles actually ‘looks like’, and importantly, what it achieves when subject to real-life commercial pressures. Given the synergies already alluded to, the ambitions of NEC would seem to find a ready parallel in the apparent successes of ‘real-life’ sociotechnical interventions. It is to one of the earliest accounts of a comprehensive organisational redesign according to sociotechnical principles, undertaken by Rice (1958) in textile mills in Ahmadabad, India, that we now turn. Here, as elsewhere, the sociotechnical redesign led to a radically different organisation which, it was argued, was jointly optimised. More than that, the ‘reorganisation was reflected in a significant and sustained improvement in mean percentage efficiency and a decrease in mean percentage damage [to goods]…the improvements were consistently maintained through-out a long period of follow up’ (Trist, 1978, p. 53). As mentioned above, the most famous example of sociotechnical design is undoubtedly that undertaken at Volvo’s Kalmar and Uddevalla car plants (e.g., Hammerström and Lansbury, 1991; Knights and McCabe, 2000; Sandberg, 1995). Whilst many commercial instantiations of sociotechnical systems theory are criticised for their limited degree of ‘technological’ change (choosing to focus instead on the altogether less expensive aspects of ‘socio’ and ‘organisational’ change; Pasmore et al., 1982) Volvo appeared to embrace the principles in an uncommonly comprehensive manner, and on a scale heretofore not yet experienced. The defining feature of the Kalmar plant’s design was a shift from a rationalistic style of hierarchical organisation to one based on smaller groups. Once again, although a world away from military command and control, conceptually, this shift retains its familiarity. In Volvo’s case the change was radical. The production line, the overtly ‘technical’ in the sociotechnical equation, not to mention the emblem of industrial age organisational design itself, literally disappeared. It was
Reconsidering the Network in NEC
23
replaced by autonomous group-work undertaken by well-qualified personnel, ‘advanced automation in the handling of production material; co-determination in the planning and a minimum of levels in the organisation’ (Sandberg, 1995, p. 147). In other words, a team of skilled workers undertook the final assembly of an entire car, from start to finish, using a kit of parts provided by the ‘advanced production material automation’. The team could also decide amongst themselves how that task was to be undertaken. From a systems perspective, according to Dekker’s more contemporary work on network structures (e.g., 2002), this new configuration had something of a ‘hybrid’ feel to it. In structural terms there was a mixture of hierarchical subdivision (albeit to a far lesser extent than before) and peer to peer interaction (within groups rather than everybody literally interacting with everyone else). In the language of NEC research what you had, in effect, were ‘communities of interest’, shortcuts between these communities and interactions that were ‘driven by circumstances’ (e.g., Alberts and Hayes, 2005; Granovetter, 1973; Watts and Strogatz, 1998). What we term here as a ‘hybrid structure’ in fact possesses some quite unique properties which we will explore in much more detail later in Chapter 7. For the time being it can be noted that, ‘the learning in this work organisation is impressive. Being engaged in all aspects of work makes the production comprehensible and the employees become, as part of their job, involved in the customer’s demands and in striving after constant improvement. Work intensity is high’ (Sandberg, 1995, p. 148). Another major effect of this network structure, as Trist (1978) notes, is that ‘whereas the former organisation had been maintained in a steady state only by the constant and arduous efforts of management, the new one proved to be inherently stable and self correcting’ (p. 53). To put this into the language of systems theory, the organisation started to behave like an open system, one that could achieve a steady state based on a constant throughput of inputs and outputs, and able to maintain this steady state despite considerable changes in the environment. In the language of NEC this phenomenon would be referred to as something like ‘selfsynchronization’ (e.g., Ferbrache, 2005). Whichever language it is translated into, these findings provide practical, large-scale, industrial validation of a conceptual approach to dealing with complexity. Organisational Interventions The specific mechanisms instantiated to support this open systems behaviour are varied and interconnected. Principle among them are natural task groupings that bestow a form of autonomous responsibility on the group, there was a ‘whole task’ to be undertaken and the requisite skills within the group to undertake it from beginning to end. In Volvo’s case the parts for the cars were organised as if they were kits, with each member of the team completing a proportion of the kit and the team as a whole effectively building a whole car (more or less independently of other teams). In terms of agility it was quickly observed that ‘model changes
24
Command and Control: The Sociotechnical Perspective
[…] needed less time and less costs in tools and training’ compared to a similar plant that was organised around the traditional factory principle (Sandberg, 1995, p. 149). Obviously, such teams still needed to be related to the wider system, which required someone to work at the system boundaries in order to ‘perceive what is needed of him and to take appropriate measures’ (Teram, 1991, p. 53). In command and control terms this new organisation shifts the primary task of commanders (or managers) away from processes of internal regulation to instead being more outwardly focused (Trist, 1978). Alberts and Hayes in their book ‘Power to the Edge’ describe a very similar role under the heading of a ‘Strategic Corporal’. This is someone who ‘must be able to function across a range of missions and make decisions that have implications far beyond his local responsibilities (Alberts and Hayes, 2005, p. 65). Back at Volvo, the similar post was labelled ‘lagombud’ or ‘group ombudsman’. Not only had the assembly line disappeared but so to had the role of supervisor. Instead, here was a strategic corporal ‘who relates to other groups and to the product shop manager’ (Sandberg, 1995, p. 148). This is an important conceptual difference. Under this sociotechnical paradigm managers and commanders now become a form of executive, coordinating function, ‘designing behaviours’ rather than arduously ‘scripting tasks’ (e.g., Reynolds, 1987). The Sociotechnical Legacy In order to provide a wider characterisation of the extant work in ‘classic’ sociotechnical systems theory a number of large meta-analyses have been identified. The first analysis is contained in a paper by Cummings, Molloy and Glen (1977) which reviewed 58 studies, the second in a paper by Pasmore et al., (1982) which reviewed 134 studies, and the third is by Beekun (1989) covering a further 17. Between them they provide a substantial overview of the first 38 years worth of practical experience in this domain. If this large body of work can be characterised at all then it can be done so with reference to the overwhelming predominance of positive study outcomes. There is no doubt that the combined results ‘support most of the claims that [sociotechnical] researchers have been making for three decades concerning the beneficial nature of this organisational redesign strategy’ (Beekun, 1989, p. 893). Table 2.1 is synthesised from Pasmore et al., (1982) and provides a good overview not just of the wider range of sociotechnical measures available and implemented in practice, but also an indication of how successful they were in relation to two dependent variables: attitudes (a more ‘socio’ variable which reflects the experience of those at work within such organisations) and productivity (a variable which reflects a form of ‘technical’ optimisation). The positive outcomes derived from applying the sociotechnical interventions listed in the left-hand column of Table 2.1 are striking. It is also interesting to note that by far the most common intervention is represented by structural change and a
Reconsidering the Network in NEC
Table 2.1
25
The use and effectiveness of common sociotechnical measures
Feature
Percentage of studies using feature (N=134)
Percentage of studies reporting success in terms of attitudes (Socio)
Percentage of studies reporting success in terms of productivity (technical)
Autonomous work groups
53
100
89
Skill development
40
94
91
Action group
22
93
100
Reward system
21
95
95
Self inspection
16
100
90
Technological change
16
92
60
Team approach
16
100
80
Facilitative leadership
14
100
100
Operators perform maintenance
12
100
88
Minimum critical specification
9
100
100
Feedback on performance
9
100
100
Customer interface
9
100
100
Self-supply
8
100
80
Managerial information for operators
7
100
67
Selection of peers
6
100
100
Status equalisation
4
100
100
Pay for learning
4
100
100
Peer review
3
100
100
94
87
% studies successful Source: Pasmore et al., 1982.
26
Command and Control: The Sociotechnical Perspective
work organisation based around teams. In other words, adapting the human to the system is far more common than adapting the technical to the human (53 per cent of studies for the former compared to only 16 per cent for the latter). Table 2.1 shows that not all applications of sociotechnical interventions are successful and it is right to guard against any over-enthusiasm. Indeed, it is certainly the case that the number of unsuccessful attempts that crop up in the public domain are likely to be fewer than those which are successful. Despite that, the preponderance of positive outcomes is difficult to ignore and there is certainly no similar legacy in NEC research. The only downside is that with so little in the way of variance it becomes difficult to judge the effect of any specific sociotechnical intervention (Cummings et al., 1977). Herein lies an interesting point. In systems-thinking it is not generally possible, or even desirable, to trace a generalised effect (productivity or attitudes for example) to specific causes. The point seems to be that sociotechnical principles and interventions are as systemic and equifinal as the system to which they are applied (e.g., Clegg, 2000). Like all good systems, perhaps they, too, become more than the sum of their parts? From the literature, then, it would seem that implementing an ostensibly ‘technical’ system like NEC is on a scale considerably in excess of many study domains analysed previously, at least within the so-called ‘classic’ sociotechnical school. What singles out NEC as a somewhat unique case is its distributed nature. For example, the idea of a roving ombudsman figure is perhaps something of an anathema in cases where teams are distributed nationally and even internationally. This kind of distribution, the separation of information from physical artefacts and locations, is an inherent part of the information-age itself. Some of the lessons to be learnt from the commercial arena will, of course, require further work in order to realise an equivalent in the domain of NEC. Another factor unique to NEC is the degree of non-linearity and complexity inherent in it. Despite the open systems principles created by sociotechnical systems theory the vast majority of the application domains are considerably more deterministic than the military arena, which has the unique property of entities in the environment that are not just dynamic but deliberately and adaptively trying to thwart your activities. There is a need to draw inspiration not only from successful classic sociotechnical studies but to also examine more contemporary developments which seek to move sociotechnical systems theory from its roots in the industrial age to a new information-age context. Contemporary Sociotechnical Systems Theory Faced with problems that are increasingly framed in terms of non-linearity and/or complexity, approaches such as macro ergonomics (e.g., Kleiner, 2006; Kirwan, 2000), cognitive systems engineering (e.g., Hollnagel, 1993; Hollnagel and Woods, 2005) and other nascent systems based developments attest to a growing
Reconsidering the Network in NEC
27
shift in human factors methods and modes of thought. As for organisations, the shift here was not due to the inherent lack of humanism but simply their ‘inability to adapt to rapid change’ (Toffler, 1981, p. 135/6). As for SST itself, British readers with a passing knowledge of industrial relations history will probably question the paradox that in the 1970s the National Union of Mineworkers voted for strike action. Their primary concern was pay but it was also to do with the hand-got method described above by Trist and Bamforth. One of the major complaints was that the working conditions were ‘un-mechanized’ (Turner, 2008). Moving forward in time to November 1992 and what most traditional texts on organisational design will tend not to mention is that the most famous sociotechnical case study of all, the Volvo Kalmar plant, is now closed. This story of apparent rise and fall might seem to cast sociotechnical systems in a poor light, as representing a slightly Luddite approach to systems design, one that has had its day and is no longer relevant to modern concerns. This would be a mistake. In the case of the coal-mining example, the sociotechnical approach did not advocate the primitive hand-got method as such; rather, it recognized that the work organisation had resulted from the co-evolution of people and environment and that large scale mechanisation represented a step change which did not necessarily recognise the favourable aspects of this hard-won experience. Sociotechnical systems are not anti-technology. As for the Volvo Kalmar plant, the reasons for its closure were considerably more obtuse. On balance, the closure of the plant was not due to a failure of the sociotechnical paradigm. Instead, it was more to do with the same shifting paradigm which favoured NEC in its early vertically integrated shock and awe form, and the resurgence of neo-Taylorism inspired by the manufacturing excellence then evident in Japan (Dankbaar, 1993). This shift coincides with the so-called ‘information revolution’ and the increasingly inventive ways in which computing and networked technology could be deployed to make complex organisations behave rationally in the face of complexity. There is no doubt that the subsequent character of sociotechnical research has been affected. Certainly, the days of ambitious large scale implementations of sociotechnical principles have largely given way to work of a smaller and somewhat more selfeffacing nature, some of which is surveyed below. Current sociotechnical thinking does, however, share with NEC an interest in the opportunities and issues raised by information technology and the internet. SST is itself shifting as evidenced by work of a more contemporary nature. Consider that a central focus of modern organisations, NEC included, is to be highly responsive to the needs of the recipients of the services which the organisation dispenses; that is to say organisations should be able to learn, the quicker and more adaptively the better (Adler and Docherty, 1998). Hirschhorn et al. (2001) refer to the idea as ‘mass-customization’ in which the real value of joint optimisation is not in the production and dissemination of ‘things’ (e.g., physical goods or ‘actions’) but in the production and dissemination of information (e.g., informational commodities like ‘effects’). They present the data shown in Table 2.2 to illustrate the conceptual shift.
Command and Control: The Sociotechnical Perspective
28
Table 2.2
Comparison of sociotechnical contexts
Focus of sociotechnical systems is on:
Focus of NEC is on:
Mass production
Mass customisation
Minimising down time
Minimising learning time
Producing product
Producing information
Maintaining a steady state
Finding information
Performing work sequentially during a defined ‘run’
Performing work continuously and adaptively
Source: Adapted from Hirschhorn et al., 2001, p. 249.
Table 2.2 clearly shows that an emphasis on traditional contexts and the micro level of organisational design would diminish the effectiveness of SST in modern settings, hence the currently expanding array of contemporary sociotechnical concepts (e.g., Heller, 1997, p. 606). Examples of these include ‘open sociotechnical systems’ (e.g., Beuscart, 2005) and sociotechnical capital (e.g., Kazmar, 2006) to name but two. The term open sociotechnical systems is an apparent contradiction in terms but actually refers to the nature of the group and to its flexible open membership. The shift is from relatively enduring semi-autonomous groups towards comparatively transient ad hoc groups. Sociotechnical capital is drawn from research in the world of the internet and a growing fascination with emergent phenomena that arises from ‘mass collaboration’ (e.g., Tapscott and Williams, 2007) and online communities. Sociotechnical capital deals with the formation and regulation of such groups and the characteristics of network systems required to support them (e.g., Resnick, 2002). Scacchi (2004) provides a contemporary summary of future research directions which, if anything, seem to be aligning the world of sociotechnical systems theory ever more closely with the concerns of NEC. They are as follows: •
•
‘First, the focus of [SST] design research is evolving towards attention to [sociotechnical interaction networks (STINs)] of people, resources, organisational policies, and institutional rules that embed and surround an information system’ (p. 5). The question of where socio and technical boundaries lie is becoming ambiguous, as is ‘who’ the user of a system is (and the panoply of values, perspectives, demands etc., that they require and/or impose). Second is ‘recognition that a large set of information systems in complex organisational settings generally have user requirements that are situated in space (organisational, resource configuration, markets, and social worlds) and time (immediate, near-term, and long term), meaning that user requirements are continuously evolving, rather than fixed’ (p. 6).
Reconsidering the Network in NEC
•
•
29
This prompts a need to decide how to access and respond to these shifting informational requirements and is reflection of similar powerful trends in Systems Engineering. Third, it is unclear how best to ‘visualize, represent, or depict (via text, graphics, or multi-media) a sociotechnical system’ (p. 6). So-called ‘Rich Pictures’ (e.g., Monk and Howard, 1998), social network diagrams (e.g., Driskell and Mullen, 2005), soft systems methodologies (e.g., Checkland and Poulter, 2006) and the Event Analysis for Systemic Teamwork methodology (e.g., Stanton, Baber and Harris, 2007; Walker et al., 2006) already reflect work underway to try and address this issue. Later chapters contribute even more directly to this aim. Fourth, ‘the practice of the design of sociotechnical systems will evolve away from prescriptive remedies to embodied and collective work practices’ (p. 7). An example is given of ‘free/open source software development projects or communities. In this sociotechnical world, the boundary between software system developers and users is highly blurred, highly permeable, or non-existent’ (p. 6–7). For example, in adapting and modifying NEC to suit local needs and preferences (as is inevitably the case; e.g., Verrall, 2006) users of NEC play as much a part in the design of the final system/ organisation as the designers themselves; whether they like it or not.
Clearly this is not an exhaustive list but it is, we believe, a characterisation, one that is expressive of the changing nature of SST, the changing nature of the systems (and their boundaries) and the changing nature of the environment to which they are applied. Summary Despite its ubiquity within human factors literature the term ‘sociotechnical’ is clearly much more than merely a buzzword. It is a set of explicit concepts, inspired by general systems theory, aimed at jointly optimising people, technology, organisations and all manner of other systemic elements. This introductory review has highlighted a key set of basic sociotechnical principles (responsible autonomy, adaptability and meaningfulness of tasks) which seem to offer favourable initial conditions for effective NEC systems. Sociotechnical principles create shared awareness (through peer to peer interaction) and agility (through effects based operations, semi-autonomous groups and increased tempo) and self-synchronisation (by joint optimisation and synergy). As Table 2.3 illustrates, the extent of overlap between the classic concept of SST and the new command and control paradigm heralded by NEC is clearly manifest. The sociotechnical approach challenges the dominant techno-centric viewpoint whereby the ‘Network’ in NEC is seen in terms of merely data and communication networks. Instead, SST speaks towards the optimum blend of social and technical
Command and Control: The Sociotechnical Perspective
30
Table 2.3
Comparison of concepts: NEC versus Sociotechnical Systems
NEC concepts
Sociotechnical concepts
Agility and tempo
Adaptability
Effects-based operations
Minimal critical specification
Peer-to-peer interaction and ad-hoc groups
Semi-autonomous work groups
Self-synchronisation
All of the above combined with whole and meaningful tasks
networks. It also refines the notions of shared awareness (the extent to which everybody needs to know everything), peer-to-peer interaction (semi-autonomous groups are a particular form of this) and, at a more fundamental level, represents a conceptual response for dealing with complexity (which is shifted from a global level of complexity to a local view of complexity which semi-autonomous groups respond to faster and more adaptively). Perhaps above all the sociotechnical approach is innately human-centred. It is concerned as much with the optimisation of effectiveness as it is with the experiences of people working within command and control organisations. Indeed, the performance of the system and the experience of those at work within it are mutually reinforcing and complementary. Arguably, the most adaptable components of all within NEC are the human actors. Sociotechnical theory, therefore, brings with it a humanistic value base and set of non-Tayloristic assumptions. It is not being offered as a panacea, but SST seems to offer considerable promise in terms of at least creating the conditions for cohesive, expert, flexible teams that relate well to a wider system. All of this would be mere conjecture were it not for SST’s 60 year legacy of applying open systems principles to commercial organisations. Whether, and by what means, the same positive outcomes could be realised in the field of NEC is something that the following chapters go on to deal with.
Chapter 3
Some Effects of Certain Communications Patterns on Group Performance Aims of the Chapter The sociotechnical perspective elucidated in the previous chapter naturally obliges us to recognise the role of humans in NEC, yet ironically this is often precisely the aspect omitted, or at least grossly over-simplified, in the design of such systems (e.g., Stanton et al., 2009). All too often, what seems to have occurred is that computer models and simulations have replaced the lab rats that researchers used to use in order to try and understand human behaviour. Harold J. Leavitt’s 1951 paper ‘Some effects of certain communication patterns on group performance’ provided a salutary lesson on the effect that different organisational structures have on human role incumbents – moreover, that it is the humans in the system who bestow upon it the agility, tempo and self-synchronisation that is desired. From this and other work in social network theory has sprung a considerable body of research which today is routinely applied to the problems of NEC. From a sociotechnical perspective, however, there is still a gap. In this chapter we present the results of a replication of Leavitt’s pioneering study, updated and refreshed for the information-age and related to the specific concerns of NEC. The task here is not to convince anyone of the validity or legacy attached to Leavitt’s foundational work (for an overview of where it fits into the wider pantheon of social network research see, for example, Martino and Spoto; 2006); instead, the goal is to put the human back into NEC and to show, via an experimental demonstration, that the principles of sociotechnical systems and of joint optimisation still apply. Whereas the previous chapter attempted to build a conceptual bridgehead between the world of sociotechnical systems theory and NEC, this chapter attempts to build an experimental one. Introduction Background and Context So what actually happens when humans are set to work within the set of organisational boundaries and constraints represented by NEC? Some insight into this question is provided by looking into a live-NEC field trial where the resultant behaviour of the system was not entirely what was expected (Verall, 2006). A substantial and wide-
32
Command and Control: The Sociotechnical Perspective
ranging human-factors analysis revealed that this particular socio-technical system, over time, became progressively more preoccupied with the ‘means’ to achieve a given end rather than the massing of objectives or ‘end states’, which is what is normally expected from NEC; indeed, what was expected in this case. Similarly, despite the provision of a networked information infrastructure, individuals and teams either used it in unpredictable ways or else adopted more familiar and presumably easier methods of working. To paraphrase Clegg (2000), what was witnessed here was ‘people interpreting the system, amending it, massaging it and making such adjustments as they saw fit and/or were able to undertake’ (p. 467). This particular case study example was ultimately successful in what it was trying to achieve which, in turn, creates a different interpretation of the results. In other words, is this apparent lack of control necessarily a bad thing? Drawing from the world whence NEC is often assumed to have come, namely the internet and other similar networked technologies, it is possible to discern several powerful trends in which this form of human adaptability, far from being commanded and controlled out of existence, is instead actively exploited. From the sublime (e.g., the Human Genome Project) to the ridiculous (e.g., Wikipedia), these are both networked, highly distributed, systems embodying the diffuse non-linear causality of peers influencing peers (Kelly, 1994; Tapscott and Williams, 2006; Viegas, Wattenberg and McKeon, 2007). These are entities where the boundary between designers and users has become ‘highly blurred, highly permeable, or non-existent’ (Scacchi, 2004, p. 6–7). Furthermore, under these ‘initial conditions’ highly effective and agile forms of organisational infrastructure ‘emerge’ rather than are created. To use Toffler’s (1981) or Tapscott and William’s (2007) phraseology, the participants in the observed NEC case study behaved rather like ‘prosumers’, individuals who see ‘the ability to adapt, massage, cajole and generally “hack” a new technology as a birthright’ (p. 32). In the human factors world Shorrock and Straeter (2006) remind us that people are still needed in complex command and control systems for precisely this reason. Moreover, this sort of human adaptability is inevitable. Therefore, perhaps a more useful way to look at NEC is not to see it as an end product or an entity that ‘is’ something, but rather as a process, something that ‘becomes’ (e.g., Houghton et al., 2006). It seems possible to go even further, to argue that an alternative conception of NEC is not something that can be called a finished article but rather as the set of initial conditions from which the most adaptable component of all, the humans in the system, create the end product most useful for their particular set of circumstances. In this we see clearly why Leavitt and others have placed such an emphasis on humans within organisations rather than lab rats within experiments or simulated agents within computer models (all of which have their place, of course). The Leavitt Study Leavitt’s 1951 study took this simple expedient, that organisational structure will in part determine human behaviour, and subjected it to a direct empirical test. Leavitt derived four distinct organisational structures, or network archetypes,
Some Effects of Certain Communications Patterns on Group Performance
33
based on small networks (of five agents). Leavitt stated that if ‘two patterns [or structures] cannot be “bent” into the same shape without breaking a link, they are different patterns’ (1951, p. 39). What emerged were the ‘circle’, ‘chain’, ‘y’ and ‘star’ networks shown in Figure 3.1. In Leavitt’s paper, the networks were populated with five subjects who were furnished with a simple problem solving task in a paradigm very similar, incidentally, to Moffat and Manso’s ELICIT concept (2008; Leweling and Nissen, 2007; Ruddy, 2007, etc.). The subjects sat in adjoining cubicles with the type of organisational structure (henceforth referred to as a ‘network’) being instigated through the provision of various open and closed letterboxes in the dividing walls. The subjects each had a set of symbols and the task was to communicate with written notes, passed through the letterboxes, what the symbol common to all subjects was. One hundred people took part in the study, split into twenty groups of five. Each group was subject to fifteen trials on each of the four network archetypes. A wide range of measures were extracted enabling both the socio and technical aspects of performance associated with each network to be derived. A clear performance dichotomy was observed: ‘the circle [network], one extreme, is active, leaderless, unorganized, erratic, and yet is enjoyed by its members. The [star network], at the other extreme, is less active, has a distinct leader, is well and stably organized, is less erratic, and yet is unsatisfying to most of its members.’ (p. 46)
Clearly, then, ‘certain communication patterns’ did indeed yield ‘some effects on group performance’. These effects prove to be of a systematic nature when a
Chain Y
Figure 3.1
Wheel
Leavitt’s network archetypes (1951, p. 39)
Circle
Command and Control: The Sociotechnical Perspective
34
further categorisation is applied. Leavitt’s Star, Chain and Y networks map, to some extent at least, onto notions of ‘classic’ hierarchical command and control. The Circle network, on the other hand, is conceptually closer to NEC. Thus in broad terms it can be seen that NEC leans more towards the ‘active, leaderless, satisfying’ and jointly optimised end of the performance spectrum compared to the technically optimised ‘stable, organised and unsatisfying’ characteristics of classic C2. This is a very crude analogy but it is not as tenuous as it might appear. Chapter 7 shows that this mapping is in fact well founded. Table 3.1, in the meantime, provides a more detailed comparison of the results obtained. Table 3.1 Network
Overall
Fastest trial
Errors (X)
N of comms. (X)
Cohesion/ enjoyment
Circle
‘showed no consistent operational organisation’ p. 42.
50.4s
3.4
419
82/100
Y
‘operated so as to give the most central position […] complete decisionmaking authority. […] Organisation for the Y evolved a little more slowly than for the [star], but, once achieved, it was just as stable’ p. 42.
35.4s
0.8
221
73/100
Star
‘The peripheral men funnelled information to the centre where an answer decision was made and the answer sent out. This organisation had usually evolved by the fourth or fifth trial and remained in use throughout’ p. 42.
32s
1.2
215.2
55.5/100
Chain
‘information was usually funnelled in from both ends […] whence the answer was sent out in both directions. The organisation was slower in emerging than the Y’s or the [Star’s], but consistent once reached’ p. 42.
53.2s
1.8
277.4
74/100
NEC Classic C2
Performance characteristics of Leavitt’s (1951) network archetypes. Shading denotes best performance in that category
Some Effects of Certain Communications Patterns on Group Performance
35
From Table 3.1 it emerges that whilst the classic C2 networks (the Star, Chain and ‘Y’) show good performance in terms of metrics like error and task completion time, as Chapter 2 hinted at, these are no longer the only criteria for judging organisational success. It could be argued that some of the attributes shown by the active, leaderless and seemingly more disorganised Circle archetype may be equally desirable in achieving a better balance between conflicting and dynamic requirements. Leavitt’s study is not ideally designed to show this as it employs what can be regarded as a very deterministic kind of task, one that would tend to favour the classic C2 networks. This gives us cause to wonder what the results would be if the desired outcomes were not error or task completion time but innovation, adaptability and emergence. In other words, how would a circle or NEC-like organisation perform on an altogether less deterministic task? It is to this that the discussion now turns. Experimental Issues The purpose of this chapter is to take the anecdotal evidence observed in the field and try to recreate, if not the exact situation, then at least an experimental analogue in the laboratory using a paradigm very similar to Leavitt’s study. The advantage of lab-based research, of course, is the degree of control that can be imposed, control that was almost entirely lacking in the field study alluded to at the beginning of the chapter. However, caution needs to be exercised. Paradoxically, too much control could conceivably prevent the emergence of the adaptive behaviour for which explanations are sought. Experimental control conflicts with ecological validity which suggests that for these types of problems perhaps a different approach to experimental design needs to be adopted. In the present study what might be referred to as a classic hierarchical C2 organisation was created within a simulated environment, then pitted against a peer-to-peer NEC counterpart, both of which contained live actors (as per Leavitt’s study) who had to operate within a complex, dynamic and adaptive scenario (unlike Leavitt’s study). Both conditions thus represent frameworks that people undertake a common task within, but different constraints apply to the different conditions. For example, there is relatively little in the way of rigid specification of experimental procedures in the NEC condition (the focus is on outcomes or ‘effects’ rather than the actions required to bring them about). In the NEC condition the technological infrastructure is also configured to facilitate peer-to-peer interaction, should the human role incumbents choose to make use of it. The opposite is true for the classic C2 condition. Here there is a high degree of ‘scripting of tasks’ and a more constrained technological infrastructure within which this occurs. The research question, therefore, shifts slightly. It focuses on the performance of the incumbents in the different organisations and how that performance changes as they adapt to their context. In a break from traditional humancentred approaches whereby the interaction (and subsequent representations)
Command and Control: The Sociotechnical Perspective
36
is generally static (Lee, 2001; Woods and Dekker, 2000) we, like Leavitt to some extent, assume it to be dynamic. Apart from Leavitt’s study there is a good basis for taking this position. Patrick, James and Ahmed (2006), for example, recognise the ‘unfolding’ nature of command and control. They state that, ‘A critical feature of command and control in safety critical systems is not only the dynamically evolving situation or state of the plant but also the fluctuating responsibilities, goals and interaction of team members’ (p. 1396). Our experimental design needs to take this inviolable aspect of command and control into account. Hypotheses Combining Leavitt’s earlier findings with the rubric of sociotechnical systems theory, along with what has been observed of live-NEC so far, the following exploratory hypotheses are put forward: •
• • • •
•
Hypothesis #1 – Differences among patterns: Leavitt’s findings lead us to expect that the C2 network will favour the commander and will develop in such a way as to maintain this pattern. Sociotechnical systems theory leads us to expect from the NEC network a more adaptive network that does not favour to the same extent one individual role incumbent. Hypothesis #2 – Time: Leavitt’s paper leads us to anticipate the classic C2 network starting off performing the task slower than the NEC network, but catching it up and overtaking it. Hypothesis #3 – Messages: Leavitt’s paper leads us to expect that the intensity of communications in the NEC condition will be higher than that for the classic C2 condition. Hypothesis #4 – Errors: Overall, Leavitt’s findings lead us to anticipate better error performance from the NEC condition as compared to the classic C2 condition. Hypothesis #5 – Questionnaire Results: Leavitt’s findings, combined with the insights provided by sociotechnical systems theory, lead us to anticipate the NEC condition leading to a better subjective experience for those at work within it as compared to the classic C2 condition. Hypothesis #6 – Message Analysis: Leavitt’s findings lead us to expect a qualitatively different type of communication to occur between NEC and classic C2 conditions. Specifically, that the interactions will be informationally richer for the NEC condition.
The insights provided by sociotechnical systems theory, combined with the observations of live-NEC and the experimental results from Leavitt (1951), lead us to anticipate better initial conditions for more effective adaptation under NEC conditions. The ‘model’, however, needs to be run in order to find out.
Some Effects of Certain Communications Patterns on Group Performance
37
Method Design The experimental task is based around a simplified ‘Military Operations in Urban Terrain’ (MOUT) game called ‘Safe houses’. The game creates a dual-task paradigm. The first task involves a commander managing two live ‘fire teams’ as they negotiate an urban environment en route to a ‘safehouse’ (note that the experiment took place in a civilian area; fire teams were not actually armed). The second task involves the commander managing the activities of ten further simulated fire teams within a much wider area of operations. The two tasks interact such that success in one does not necessarily connote success overall. It falls to the commander to effectively balance task demands under the independent, between subjects variable of command and control ‘type’, which has two levels: NEC and C2. The study is longitudinal in nature. The two teams (NEC and C2) separately undertook a total of 30 iterations through the same dynamic task paradigm and a simple form of time series analysis was employed to reveal the underlying, highlevel ‘adaptive model’ embedded in the data. Participant matching, training and randomisation were employed to control for individual differences, order effects and task artefacts. Several ‘socio’ aspects of performance were measured. Thirty separate socialnetwork analyses were undertaken, with the network metrics so derived being subject to the same form of simple regression-based time series analysis in order to show how the pattern of communication changed over time. This combined with an analysis of communications quantity and content (also dependent variables), the latter being facilitated with a pre-defined communications taxonomy. The final ‘socio’ dependent variable was the subjective experience of the role incumbents, which was measured over seven time intervals using a modified self-report cohesion questionnaire. Several ‘technical’ aspects of performance were also measured. Good performance in these terms equates to the shortest task completion time, all en route Target Areas of Interest (TAI’s) correctly located and effected, and a high ratio of enemy to friendly agents neutralised. Participants There are five principle roles in the study, three of which were occupied by experimental participants (all aged 21). The remaining two roles were filled by the experimenters. The experimental roles are shown in Table 3.2. Materials Command and Control Microworld Figure 3.2 presents a visual representation of the command and control microworld within which the safehouses game was implemented.
38
Table 3.2
Command and Control: The Sociotechnical Perspective
Experimental participants and their roles
Role
Role purpose
Commander
The incumbent of this role was in charge of both fire teams within the primary task, providing guidance and strategy as required. They were also responsible for the larger strategic secondary task.
Fire Team Alpha
This participant was located within the live battlespace and communicated to the commander, and depending on experimental condition, the other fire team as well, by using an XDA mobile device and MSN Messenger™. The XDA device also enabled the fire team to be live-tracked and represented on the commander’s command-wall representation of the battlespace.
Fire Team Bravo
This participant had the same role and capabilities as Fire Team Alpha.
NEC System Operator (Performed by member of experimental team)
The NEC System Operator dealt with the experimental aspects of the commander’s primary task (i.e., managing the fire teams) as well as the NEC system itself. The NEC System Operator effectively ‘drove’ the NEC command-wall system, receiving requests to add/append data to the live maps from the commander and helping them to use the system themselves. The system operator also supplied strategic injects according to pre-set rules. In the NEC condition, the experimenter also provided situational updates to all team members (ensuring that ‘everyone’ knew ‘everything’).
Enemy (Performed by member of experimental team)
This individual, like the NEC system operator, was another member of the experimental team but this time physically isolated from the battlespace, the commander and the NEC system operator. They were in charge of playing the commander in the secondary task, thus they controlled enemy actions in a ‘Wizard of Oz’ fashion.
The system operator and commander were co-located in the laboratory. Both sat with a clear view of the so-called ‘command wall’ which was a large projection screen displaying a modified Google Earth™ representation of the battlespace. This visualisation was annotated with the live position of the fire teams (via GPS) as represented by an appropriate icon. This visualisation was supplemented by a separate large screen displaying a ‘planning window’ which contained a simple map-based representation of the battlespace environment with a grid-square coordinate system superimposed over the top of it. The planning window allowed the system operator and commander to add, delete
Some Effects of Certain Communications Patterns on Group Performance Classic C2 Communication Links Additional Links for NEC Condition
39
Command Centre
System Operator
Commander
Alpha
Enemy Bravo Next generation XDA devices
Areas of Operations #1
Figure 3.2
Command and control microworld
and move objects as required by the primary and secondary tasks. Any such changes were instantaneously represented on the main visualisation display. The planning window was populated by the experimenter (before the respective experimental condition commenced) with all the required TAI’s and enemy icons positioned according to a preset template which was randomly selected for that trial. The commander and system operator had separate work stations and their own computer, and communicated purely through text based means (using MSN Messenger™). Both fire teams (Alpha and Bravo) were located in the physical battlespace, away from the co-located commander and system operator. The fire teams carried an XDA mobile-phone device and it is this that permitted them to be live-tracked using GPS. The XDA device also allowed each fire team to communicate with each other as well as, although depending on experimental condition, the commander in the control centre. Again, all such interactions were performed textually via MSN Messenger™. Fire teams could also use the XDA to add icons into their own miniature version of the digital map which would then simultaneously appear on each others screens as well as the main visualisation window in the command centre. The miniature digital map shown on the fire teams’ XDA screens was also used for navigation purposes.
40
Command and Control: The Sociotechnical Perspective
Command and Command Paradigms The NATO SAS-050 model of command and control (NATO, 2007) was used to design Leavitt-like networks and a set of organisational constraints that combined to enable appropriate NEC and C2 characteristics to be created, as shown in Table 3.3. ‘Safe houses’ Game (Primary Task) The safe-houses game was inspired by a military training guide (MoD, 2005) and involved the fire team choosing and negotiating a route through an urban battlespace in order to correctly locate and effect a so-called safe house. In addition to this they had to deal with a number of Target Areas of Interest (TAIs) en route, with each team having to collaborate by providing cover for the other. Both fire teams start from the same location. The location of the safe house, the final destination, is randomised for each trial but with distance (and potential task time) from the start point controlled. The area of operations for the primary task is scattered with numerous TAIs so that an equal number of them will be encountered en route to the safehouse regardless of the route chosen to get there. Firstly, each of these ‘en route’ TAIs had to be correctly located by one of the fire teams. Correct location of a TAI is judged to have occurred when the fire team takes up position at the same grid coordinates as the TAI. This position is shown Table 3.3
NATO SAS-050 Model of Command and Control was used to design NEC and C2 command organisations with the appropriate characteristics NEC
C2
DISTRIBUTION OF INFORMATION
BROAD: fire teams are provided with regular situation updates from the system operator in addition to being able to interact directly with their counterparts. ‘Everyone’ knows ‘everything’.
TIGHT: the commander is the only individual with an overall view of the situation. The fire teams’ had a local view of their immediate location but in all other respects worked in isolation. ‘Everyone does not know everything’.
PATTERNS OF INTERACTION
DISTRIBUTED: all team member roles could speak to each other independently (there was no communications hierarchy).
HIERARCHIAL: the fire teams could speak to the commander but not directly to each other.
ALLOCATION OF DECISION RIGHTS
PEER TO PEER: collaborative working encouraged and facilitated by effects based instructions and peer-to-peer communications infrastructure.
UNITARY: autonomy, authority and discretion rested with the commander only.
Some Effects of Certain Communications Patterns on Group Performance
41
on the XDA device’s digital map display. Missing out a TAI by failure to locate it will result in the offending fire team being removed from the mission and having to return to the start point for the remainder of the trial. Secondly, assuming the en route TAI has been correctly located by the fire team, who will now be standing on the requisite grid square coordinates, the fire team then has to ‘effect’ it in order to make it safe for the other fire team to continue on their route. Although the location of the TAI is known by the fire team and commander a priori, what is not known is what form it actually takes and the most appropriate way to ‘effect’ it. This can only be judged by the fire team who are on the ground and are able to make that assessment based on a number of simple local characteristics. These are as follows: •
• •
If the TAI coordinates correspond to the side of a building over three storeys high then a ‘yellow effect’ will neutralise it which is signified by the relevant fire team using their XDA to place a yellow icon on the appropriate grid coordinate. If the TAI is located on the side of a building less than three storeys high, then a ‘blue effect’ will neutralise it. This is signified by a blue icon being placed. If the TAI is located in the middle of a thoroughfare then a ‘pink effect’ will neutralise it, signified by a pink icon.
After confirmation that this information has been received, the relevant fire team will hold in this position, providing cover for the other fire team as they make their way to the next TAI. This ‘leapfrogging’ effect continues until, finally, the safehouse itself is located and affected in exactly the same way. It should be pointed out that in the NEC condition this leap-frogging is facilitated by the fire teams’ being able to communicate directly with each other via their XDAs; in the C2 condition, however, communication (and instructions) have to pass through and/or come from the commander. In order to further encourage the need for communication and interaction there is also a degree of built-in ambiguity in the positional data. This means that part of the adaptive process of the entire team is to figure out ‘work-arounds’ and modes of operation that enable these ambiguities to be resolved in whatever way is found to be most efficient. The need for good time and accuracy performance is embedded in the scenario by two simple game-play expedients. As mentioned before, if the wrong location is chosen or the TAI is ignored then the fire team allocated to it fails the mission and has to sit out the remainder of the trial. If the right location but wrong effect is applied then the fire team’s attrition score, which starts at five and denotes ‘full strength’, is decremented. If the attrition score for a particular fire team reaches zero, their effectiveness is neutralised and they cannot continue the task. The attrition score is not just affected by accuracy but also speed and time. Five timeactivated attrition-injects occur throughout the fifteen minutes allotted to the trial;
42
Command and Control: The Sociotechnical Perspective
these cause both fire team’s attrition scores to be decremented the longer they spend in the battlespace. What appears to be a relatively complex set of rules becomes considerably simplified as far as the experimental participants are concerned. The system operator (who is a member of the experimental team) undertakes all game-play management tasks such as maintaining the formal record of location accuracy, whether the right effect has been applied, the attrition score, enactment of the time-based attrition-injects, and communicates all of this to the commander as required. Secondary Task (Commander) Whilst the first task concerns the activities of a live-fire team in a live environment, the secondary task of the commander is simulated and occurs in a much wider area of operations (AO). The boundaries of this are six miles in either direction from the AO of the first task. This larger strategic mission relies on the commander playing a competitive game against a simulated enemy played by the second experimenter, who is acting in a wizard of oz fashion. Thus ‘live’ (small AO) and ‘simulated’ (large AO) elements are played simultaneously on the same NEC system. So, whilst the live fire team progress towards their primary target, within their smaller AO, wider enemy activity is taking place all around which, if permitted to continue, will eventually impinge on the primary task. Such impingement, under certain specified conditions, will mean that the primary task fails regardless of the efficacy of the live-fire teams. The onus is thus on the commander to manage both tasks effectively. In the commander’s wider area of operations there are fifteen enemy elements/ icons dispersed randomly around the environment. These icons are placed into the main visualisation window of the NEC system according to a random script for that trial, by the experimenter acting in the ‘enemy’ role. Enemy icons, due to the asymmetry of most MOUT-type situations, are free to move across the battlespace at will with no area being restricted to them. On the other side, the commander has 25 friendly-force icons/elements all massed in a defined ‘green zone’; this is their starting position for the game. Although far more numerous, the friendly icons are not permitted to enter predefined ‘sensitive areas’. Apart from these constraints, the game plays like a virtual game of draughts in which the grid square coordinate system of the map serves as a form of draughts board. Only one icon can be moved, one square at a time (in any direction) in enemy reaction, friendly counter action, enemy counter re-action, and so on in sequence. If the friendly icon enters a grid square occupied by an enemy icon then the friendly icon wins. If the enemy icon enters a grid square occupied by a friendly icon, the enemy wins. The commander controls his friendly pieces through textual means (again with MSN Messenger™) by telling the experimenter acting in the enemy role which coordinates he would like his particular friendly piece to move to. The experimenter playing the role of enemy updates the icons/
Some Effects of Certain Communications Patterns on Group Performance
43
map accordingly and keeps a record of the secondary task score, which is the ratio of enemy/friendly pieces removed from the game. Cohesion Questionnaire A modified version of the Combat Platoon Cohesion Questionnaire (Siebold and Kelly, 1988) was deployed in order to provide some measure of the role incumbents’ subjective experience of the two structures (NEC and C2). The questionnaire was shortened to fifteen items, all of which required a response along a 7-point Likert scale. A score of one equates to the maximum score on any given cohesion question, 7 the minimum. The individual question items, and the corresponding mapping to Hackman and Oldman’s (1980) core job-characteristics and the sociotechnical concept of ‘meaningfulness of tasks’, are shown below in Table 3.4. Table 3.4
Question items drawn from the Platoon Cohesion Questionnaire (Siebold and Kelly, 1988) mapped onto Hackman and Oldman’s (1980) core job characteristics to create a much simplified assessment method Core job characteristic
Question item Skill
Task identity
Task significance
Autonomy
These items rated from ‘not at all important’ to ‘extremely important’ Loyalty to the team Taking responsibility for their actions and decisions Accomplishing all assigned tasks to the best of their ability Commitment to working as members of a team Dedication to learning their job and doing it well Personal drive to succeed in the tasks Taking responsibility to ensure the job gets done The team can trust one another Feel close to other team members
Feedback
44
Table 3.4
Command and Control: The Sociotechnical Perspective
Concluded Core job characteristic
Question item Skill
Task identity
Task significance
Autonomy
Feedback
These items rated from ‘very much/well/always etc’ to ‘very little/poorly/not at all etc’ How well do the members in your team work together? To what extent do team members help each other to get the job done? To what extent do team members encourage each other to succeed? Do the members of your team work hard to get things done? The chain of command works well? Everyone is well informed about what is going on?
Procedure Training Phase (Day #1) The aims and objectives of the experimental task were introduced to the external participants along with health and safety preliminaries and informed consent. Detailed instructions on the task were then provided to all participants, supplemented with demonstrations and hands-on examples. The experimenter then used the prepopulated command-wall to begin the first full trial which was identical in all respects to the experimental trials but serves initially as a practice (both teams are measured subsequently as an internal check on concordance between them). Experimental Phase (Day #2–8) With the participants familiar with the task paradigm, the teams begin to undertake the repeated iterations of it. Issues and questions are dealt with before the trial starts and during it if required. The task gets underway and the participants interact in the manner prescribed by the organisational type they are currently working
Some Effects of Certain Communications Patterns on Group Performance
45
within. They have a maximum time of fifteen minutes within which to complete the task. Results To briefly recap, two teams took part in a simulated MOUT mission over 30 successive iterations. The analysis focuses on how the different constraints of NEC and C2 influenced the direction of team adaptation and performance over time. In broad terms it is hypothesised that NEC will provides better conditions for adaptation to the complex dynamic experimental task. A simple form of time series analysis will be deployed in order to uncover the underlying theory behind the data and thus put this supposition to the test. The results that follow are compared to Leavitt’s original findings and presented under that paper’s original sub-headings: Differences Among Patterns Of the network structures more pertinent to classic C2 Leavitt (1951) writes: ‘The star operated in the same way […] The peripheral men funnelled information to the centre where an answer decision was made and the answer sent out. […] The Y operated so as to give the most-central position […] complete decision-making authority. […] In the chain information was usually funnelled in from both ends […] whence the answer was sent out in both directions’ (p. 42). Of the network structure most pertinent to NEC the following is noted: ‘The circle showed no consistent operational organisation’ (Leavitt, 1951, p. 42). Our findings show the following. The structure of communications (as in who can speak to whom) is less constrained in the NEC condition than it is in the C2 condition. In the NEC condition everyone can speak to anyone (if they want to), whereas in the C2 condition the structure of communications is fixed and all communications have to pass through the commander. Social Network Analysis (SNA) lends itself well to examining structural determinates of communication like these. Briefly, a social network can be plotted by examining ‘who’ is communicating to ‘whom’. The presence of a communication establishes a link between any actors that are communicating with each other, creating its own structure in the form of a social network. The network can then be examined mathematically to identify key characteristics such as sociometric status (who, based on their degree of interconnection, can be regarded as a key agent) and density (how densely interconnected the network as a whole is). In the present case, density and sociometric status did not behave in a way that made the linear regression method of time series analysis tenable. In terms of both these measures the regression diagnostics showed only weak (and non-significant) associations between them and trial interval. Likewise, the regression ANOVAs failed to detect a meaningful and/or significant effect in terms of the data behaving
46
Command and Control: The Sociotechnical Perspective
linearly. This form of modelling is therefore substituted for a more conventional cross-sectional analysis. The simplicity and constraints imposed on both the NEC and C2 social networks are clearly evident in the behaviour of the density metric. In both cases (albeit during different trial intervals) the density figure only deviated from a fixed point on three occasions. This behaviour makes non-parametric tests appropriate. The results show that the mean density figure in the NEC condition (M = 0.97, SD = 0.1) was somewhat higher than the C2 condition (M = 0.61, SD = 0.19). A Mann-Whitney U test goes on to suggest that the NEC social network is significantly more dense (U = 40.5; exact p < 0.01) than the C2 network, meaning that the opportunity for peer-to-peer collaboration was taken by the study participants who, remember, were under no obligation to use it. The approximate effect size of this finding is rbis = 0.25, a relatively small effect yet one consistent with the small size and simple characteristics of the networks under analysis. The findings for sociometric status are presented in Table 3.5. Table 3.5
Results of analysing sociometric status in relation to NEC and C2 conditions
Agent
Type
Mean sociometric status
Standard deviation
U statistic
Exact p
Approx effect size (rbis)
Commander
NEC
7.88
3.27
52.5
< 0.0001
0.26
C2
20.95
8.99
NEC
11.92
2.73
417
= 0.32
0.04
C2
12.02
5.39
NEC
12.13
3.29
222.5
< 0.001
0.22
C2
8.93
4.12
Alpha
Bravo
The pattern of results shown in Table 3.5 paints quite a coherent picture. As anticipated, in the C2 condition the commander has significantly elevated status compared to the NEC condition (mimicking the results observed for Leavitt’s Star, Y and Chain networks). The NEC condition, on the other hand, was not as inconsistent as Leavitt’s study might have suggested. Faced with this particular complex, dynamic and adaptive task the opportunity for peer-to-peer interaction was seized upon and exploited. This, in combination with elevated levels of communication traffic, led to higher network density and a more even spread of sociometric status among the participating agents. In sociotechnical terms this data is suggestive of a greater degree of internal leadership and autonomy, with all agents able (as in fact was the case) to communicate and participate with no single agent becoming particularly dominant.
Some Effects of Certain Communications Patterns on Group Performance
47
Task Performance Time In regard to task performance time Leavitt (1951) writes: ‘A measure of the fastest single trial of each group indicates that the star [i.e., classic C2] was considerably faster (at its fastest) than the circle [i.e., NEC]’ (p. 43). The adaption curves presented by Leavitt also trace a distinct pattern, one in which the circle/NEClike archetype starts off performing the task more quickly but the curve for the star/classic C2-like archetype eventually overtakes it. In other words, the curves cross with the result that the star/classic C2 archetype becomes quicker than the circle/NEC archetype even though it did not start out that way. Our simple time series findings are more or less identical in this regard. Both teams (C2 and NEC) were measured in terms of how long it took the live-fire team to perform their task. When this first task was complete then both tasks of the dual-task paradigm were halted. The maximum amount of time that was allowed to be spent on the task was 15 minutes (900 seconds). As one would expect, over the course of the 30 iterations both teams speeded up considerably and continued to do so for every trial. This we can consider as a linear first order effect which we can analyse using linear regression as a method of time series
Figure 3.3
Scatter plot showing the regression lines for both NEC and C2 conditions in relation to task time
48
Command and Control: The Sociotechnical Perspective
analysis. A strong association between task time and trial was obtained for both conditions (NEC r = -0.84 and C2 r = -0.85), both of which were significant at beyond the 1 per cent level. Furthermore, the regression ANOVA supports the hypotheses that this association is linear in nature for the NEC condition: F(1,28) = 64.74; p < 0.0005 and F(1,28) = 73.53; p